Michigan Law Review
Volume 105 | Issue 4
2007
Voter Identification
Spencer Overton
The George Washington University Law School
Follow this and additional works at: http://repository.law.umich.edu/mlr
Part of the Civil Rights and Discrimination Commons, Election Law Commons, and the State
and Local Government Law Commons
Recommended Citation
Spencer Overton, Voter Identification, 105 Mich. L. Rev. 631 (2007).
Available at: http://repository.law.umich.edu/mlr/vol105/iss4/1
This Article is brought to you for free and open access by the Michigan Law Review at University of Michigan Law School Scholarship Repository. It
has been accepted for inclusion in Michigan Law Review by an authorized administrator of University of Michigan Law School Scholarship Repository.
For more information, please contact [email protected].
VOTER IDENTIFICATION
Spencer Overton*
In the wake of closely contested elections, calls for laws that require voters
to present photo identification as a condition to cast a ballot have become
pervasive. Advocates tend to rely on two rhetoricaldevices: (1) anecdotes
about a couple of elections tainted by voter fraud; and (2) "common
sense" arguments that voters should produce photo identification because
identification is required to board airplanes, buy alcohol, and engage in
other activities. This Article explains the analytical shortcomings of anecdote, analogy, and intuition, and applies a cost-benefit approach generally
overlooked in election law scholarship. Rather than rushing to impose a
photo-identification requirement for voting, policymakers should instead
examine empirical data to weigh the costs and benefits of such a requirement. Existing data suggest that the number of legitimate voters who would
fail to bring photo identification to the polls is several times higher than
the number of fraudulent voters, and that a photo-identification requirement would produce political outcomes that are less reflective of the
electorate as a whole. Policymakers should await better empirical studies
before imposing potentially antidemocratic measures. Judges, in turn,
should demand statistical data to ensure that voter identification procedures are appropriately tailored to deter fraudulent voters rather than
legitimate ones and do not disproportionately exclude protected classes of
voters.
TABLE OF CONTENTS
633
638
A. Existing State Laws for Identifying Voters......................... 639
B. Photo-IdentificationRequirements to Vote ........................ 641
INTRODU CTION ......................................................................................
I. THE VOTER IDENTIFICATION LANDSCAPE ................................
*
Associate Professor of Law, The George Washington University Law School. Michael
Abramowicz, Bob Bauer, David Becker, Tom Colby, Jamie Grodsky, Paul Herrnson, David Hyman,
Michael Kang, Ellen Katz, Dan Ortiz, Leslie Overton, Richard Pildes, Peter Smith, Amanda Tyler,
Tova Wang, and Fane Wolfer read earlier drafts of this Article and provided helpful comments. The
Article also benefited from my interaction with Tom Daschle and Raul Yzaguirre in formulating our
dissent to the Commission on Federal Election Reform's photo-identification proposal; conversations with Wendy Weiser and Justin Levitt of the Brennan Center for Justice at NYU School of Law
during the drafting of our Response to the Report of the 2005 Commission on Federal Election
Reform; and discussions with Adam Cox, Heather Gerken, Michael Kang, and Daniel Tokaji during
the drafting of our letter to the U.S. Department of Justice regarding Georgia's photo-identification
law. Exchanges with Steve Carbo, David Dill, John Duffy, Chris Edley, Grant Hayden, Gracia Hillman, Ellen Katz, Orin Kerr, Bill Kovacic, Lori Minnite, Jon Molot, Peter Swire, Dan Tokaji, Clyde
Wilcox, and Brenda Wright also helped develop my thinking. Daniel Taylor provided invaluable
research assistance.
Michigan Law Review
II.
III.
[Vol. 105:631
THE SHORTCOMINGS OF ANECDOTE, ANALOGY,
AND INTUITION TO JUSTIFY PHOTO IDENTIFICATION ...............
644
A. Misleading and UnrepresentativeAnecdotes
about Voter Fraud..............................................................
B . FlawedA nalogies..............................................................
644
650
THE NEED FOR EMPIRICAL EVIDENCE TO BETTER
UNDERSTAND FRAUD AND ACCESS ..........................................
A. Toward Better Data on the Extent of Fraud......................
1. Investigations and Prosecutionsof Voter Fraud.........
2. Random Surveys of Voters ...........................................
3. Examining Death Rolls ...............................................
B. Toward Better Data on Legitimate Voters
Excluded by Photo Identification ......................................
IV.
THE LEGAL STATUS OF PHOTOIDENTIFICATION REQUIREMENTS .............................................
A. Burdening the FundamentalRight to Vote .........................
1. Assessing the Voters'Burden Relative
to the State Interest.....................................................
2. Tailoring......................................................................
B. Photo-IdentificationFees as Poll Taxes.............................
C. Abridging Voting Rights along Racial Lines .................
D. "Individual Responsibility" in the
Context of Democracy.......................................................
V.
653
653
654
655
656
657
663
664
665
667
669
670
672
PHOTO-IDENTIFICATION SUPPLEMENTS
AND ALTERNATIVES..................................................................
A. Supplements That May Enhance Voter Access ..................
1. Free Photo Identification............................................
2. Expanded Photo-IdentificationDistribution
through Mobile Buses and More Photo
Identification Offices...................................................
3. ProvisionalBallots Counted When Photo
Identification Presented..............................................
4. Election Day Registration...........................................
B. Alternatives That Allow Voters Who Lack Photo
Identification to CastBallots.............................................
1. Nonphoto Identification..............................................
2. Requiring Photo Identification at Registration
Rather than at the Polls ..............................................
3. Signature Comparison................................................
4. Affi davits .....................................................................
5. Indelible Ink ................................................................
6. Government MaintainsDigital
Picture/BiometricIThumbprint
....................................
7. Better Election AdministrationPractices...................
CONCLU SION .........................................................................................
674
674
675
676
677
677
678
678
678
679
679
679
679
680
68 1
February 20071
Voter Identification
633
INTRODUCTION
I served as a member of the Commission on Federal Election Reform, a
bipartisan, private commission tasked with proposing solutions to America's
most pressing election problems. Former Democratic President Jimmy
Carter and former Republican Secretary of State James Baker co-chaired the
twenty-one-member body,' and other commissioners included
S
2 former members of Congress, cabinet officials, and university presidents. On September
19, 2005, the "Carter-Baker Commission" released eighty-seven different
recommendations, one of which proposed that voters produce a photoidentification card as a condition to casting a ballot.3 I dissented from the
proposed photo-identification requirement, as did two other commission
members.4
The Commission's photo-identification proposal received extensive
media attention and fueled a firestorm of similar proposals across the
nation.5 Georgia, Indiana, and Missouri have adopted laws making them the
only states to prohibit citizens from casting ballots unless they produce
valid, government-issued, nonexpired photo identification,6 and bills that
tighten voter-identification requirements have been introduced recently in
1. President Carter and Secretary of State Baker had their own experiences with election
problems. President Carter led delegations that monitored elections in countries around the world,
and Secretary of State Baker led the George W. Bush campaign during the disputed Florida presidential election recount in 2000.
2. Robert Pastor, an advisor to President Carter in trips abroad to monitor elections,
organized the Carter-Baker Commission in early 2005 through the Center for Democracy and
Election Management and served as Executive Director of the Carter-Baker Commission. Former
U.S. Senate Democratic Leader Tom Daschle, former Democratic Congressman and 9/11
Commission Chair Lee Hamilton, former Republican Congresswoman Susan Molinari, and former
Republican U.S. Secretary of Commerce Robert Mosbacher were some of the other more
recognizable commissioners. For a complete list of commission members, see Commission on
Federal Election Reform-Members, http://www.amefican.edu/ia/cfer/members.htm (last visited
Sept. 25, 2006).
3. See also Robert A. Pastor, Improving the U.S. Electoral System: Lessons from Canada
and Mexico, 3 ELECTION L.J. 584, 588 (2004) (proposing a variety of election reforms, including a
photo-identification requirement to vote).
4. While several Commission members expressed strong criticisms of a photo-identification
requirement during our final Commission meeting, only three of us issued a formal dissent-former
U.S. Senator Tom Daschle, former National Council of La Raza President Raul Yzaguirre, and
myself. COMM'N ON FED. ELECTION REFORM, BUILDING CONFIDENCE IN U.S. ELECTIONS 88-89
(2005), http://www.american.edu/ia/cfer/report/full-report.pdf. Unfortunately, the three of us were
prevented from including in the Report an extensive analysis of a photo-identification requirement's
costs and benefits because of a rule limiting dissents to 250 words per commissioner. The rule was
first announced by Executive Director Robert Pastor at our final meeting.
E.g., Dan Balz, Carter-Baker Panel to Call for Voting Fixes: Election Report Urges
5.
Photo IDs, Paper Trails and Impartial Oversight, WASH. POST, Sept. 19, 2005, at A3; James Gerstenzang, Election Overhaul is Urged, L.A. TIMES, Sept. 19, 2005, at A7; Report Urges Photo IDs to
Curb Vote Fraud, CoM. APPEAL (Memphis, Tenn.), Sept. 20, 2005, at A4; The NewsHour with Jim
Lehrer,Conversation: Carterand Baker (PBS television broadcast Sept. 19, 2005).
6.
GA. CODE ANN. § 21-2-417 (Supp. 2006); IND. CODE ANN. § 3-11-8-25.1 (West Supp.
2006); Mo. ANN. STAT. § 115.427 (West Supp. 2006), invalidated by Weinschenk v. State, No. SC
88039, 2006 WL 2959284 (Mo. Oct. 16, 2006) (en banc).
Michigan Law Review
[Vol. 105:631
Congress and the majority of state legislatures.7 Polls show that eighty-one
percent of Americans favor or strongly favor requiring voters to produce
photo-identification cards before voting.' Several recommendations of the
Commission's 2001 predecessor-the Carter-Ford Commission-were enacted into law in the Help America Vote Act of 2002, 9 and hopeful photoidentification advocates repeatedly cited the 2005 Carter-Baker Commis-
sion's recommendation to bolster their proposals.'0
This Article is the first academic work to analyze photo-identification
requirements in depth, and employs an empirical cost-benefit approach to
expose the erroneous assumptions of conventional wisdom." It argues that
before jumping on the photo-identification bandwagon, policymakers should
closely examine empirical data about the magnitude of voter fraud and the
extent to which a photo-identification requirement would reduce participation by legitimate voters. While a small amount of voter fraud
hypothetically could determine a close election, the exclusion of twenty mil-
lion Americans who lack 2photo identification could erroneously skew a
larger number of elections.1
7. H.R. 4844, 109th Cong. § 2 (2006) (requiring that voters in federal elections provide
current and valid government-issued photo identification-the bill passed 228-196); S. 414, 109th
Cong. § 203(b) (2005) (proposing legislation that would require all in-person voters in federal elections to present current and valid photo identification before voting).
See, e.g., PETER HART & BILL MCINTURFF, NBC NEWS/WALL STREET JOURNAL SUR# 6062, at 13 (2006), http://online.wsj.com/public/resources/documents/
po1l20060426.pdf (finding that in April of 2006, 62% of respondents to a national poll strongly
favored the showing of photo identification before voting, 19% somewhat favored, 12% were neutral, 3% somewhat opposed, and only 4% strongly opposed).
8.
VEY,
STUDY
9. Congress adopted the 2001 Carter-Ford Commission's proposals for provisional ballots,
statewide voter registration lists, and the creation of the Election Assistance Commission. NAT'L
COMM'N ON FED. ELECTION REFORM, To ASSURE PRIDE AND CONFIDENCE IN THE ELECTORAL
PROCESS 6, 13 (2001), http://millercenter.virginia.edu/programs/natlcommissions/commission_
final-report/fullReport.pdf.
10. E.g., Jo Mannies, Measure to Require Photo IDs Stirs Outcry, ST. LOUIS POST-DISPATCH,
Feb. 12, 2006, at B I ("Thor Heame, a prominent Republican who has been pushing [photo ID]
legislation in several states ... notes that photo identification was among the recommendations of
the bipartisan Commission on Federal Election Reform...."); Dane Smith, Panel OKs Bill Requiring Citizenship Proof to Vote, STAR TRIB. (Minneapolis, Minn.), Mar. 16, 2006, at lB (explaining
that the Republican sponsor of a state bill requiring voters to show proof of citizenship noted that
the bipartisan commission recommended a photo-identification requirement).
11. A few other articles list photo-identification proposals along with a bundle of other election reforms or election law developments but do not analyze the proposals extensively. See
Developments in the Law, Voting and Democracy, 119 HARV. L. REV. 1127, 1144-54 (2006); Richard L. Hasen, Beyond the Margin of Litigation: Reforming U.S. Election Administration to Avoid
Electoral Meltdown, 62 WASH. & LEE L. REV. 937, 969-70 (2005); Pastor, supra note 3, at 588;
Publius, Securing the Integrity of American Elections: The Need for Change, 9 TEx. REV. L. & POL.
277, 288-89 (2005); see also Dan Eggen, Official's Article on Voting Law Spurs Outcry, WASH.
POST, Apr. 13, 2006, at A19 (identifying Hans von Spakovsky-a senior lawyer in the Justice Department's Civil Rights Division who played a critical role in overruling career attorneys and
approving Georgia's identification program-as "Publius," the author of the Texas Review of Law
and Politicsarticle).
12. See BRENNAN CTR. FOR JUST. AT N.YU. SCH. OF LAW & SPENCER OVERTON, RESPONSE
TO THE REPORT OF THE 2005 COMMISSION ON FEDERAL ELECTION REFORM 24 n.9 (2005) (estimating that twenty-two million voting-age citizens lack a driver's license, based on analysis of 2000
Census and 2003 Federal Highway Administration data); TASK FORCE ON THE FED. ELECTION SYS.,
February 2007]
Voter Identification
No systematic, empirical study of the magnitude of voter fraud has been
conducted at either the national level or in any state to date, 3 but the best
existing data suggests that a photo-identification requirement would do
more harm than good. An estimated 6-11% of voting-age Americans do not
possess a state-issued photo-identification card, and in states such as Wis-4
consin 78% of African-American men ages 18-24 lack a driver's license.1
By comparison, a study of 2.8 million ballots cast in 2004 in Washington
State showed only 0.0009% of the ballots involved double voting or voting
in the name of deceased individuals." If further study confirms that photoidentification requirements would deter over 6700 legitimate votes for every
single fraudulent vote prevented, a photo-identification requirement would
increase the likelihood of erroneous election outcomes.
This Article is important because political sound bites and media reports, rather than comprehensive academic analysis, have shaped the photoidentification debate. As a result, many Carter-Baker Commission members,
Justice Department officials, members of Congress, governors, state legislators, newspaper columnists, and average citizens have embraced flawed
assumptions by relying on a story or two about voter fraud. While anecdotes
about fraud are rhetorically persuasive, the narratives often contain false
information, omit critical facts, or focus on wrongdoing that a photoidentification requirement would not prevent. Even when true, anecdotes do
not reveal the frequency of similar instances of voter fraud.
The current popular debate has also relied on flawed analogies, with advocates asserting that photo-identification cards are commonly required to
curb terrorism, prevent credit card fraud, and protect minors. They do not,
however, explore why people are allowed to engage in many activities-in
To ASSURE
PANY
PRIDE AND CONFIDENCE IN THE ELECTORAL PROCESS: TASK FORCE REPORTS TO ACCOMTHE REPORT OF THE NATIONAL COMMISSION ON ELECTION REFORM ch. 6 (2001),
http://www.millercentervirginia.edu/programs/nat-commissions/commission-final-report/task-for
ce-report/taskjforcecomplete.pdf.
13.
CHANDLER DAVIDSON ET AL., REPUBLICAN BALLOT SECURITY PROGRAMS: VOTE PRO-
OR MINORITY VOTE SUPPRESSION-OR BOTH? 99 (2004), http://www.votelaw.com/
blog/blogdocs/GOP BallotSecurityPrograms.pdf; U.S. ELECTION ASSISTANCE COMM'N, ELECTION CRIMES: AN INITIAL REVIEW AND RECOMMENDATIONS FOR FUTURE STUDY 1 (2006),
TECTION
http://www.eac.gov/docs/Voter%20Fraud%20&%201ntimidation%20Report%20-POSTED.pdf
(concluding that there has never been a comprehensive, nationwide study of voter fraud based on a
review of existing literature and interviews with experts). Rather than wait twelve to eighteen
months for teams of researchers to compile and publish the extensive studies proposed in Part III,
and risk the chance that politicians in dozens of states will continue to introduce and enact photoidentification requirements that could potentially exclude millions of legitimate voters, this Article
compiles the best data currently available on voter fraud and voter access to assert that lawmakers
should place a moratorium on more restrictive voter-identification proposals until they obtain a
better empirical understanding of the extent and nature of voter fraud.
14.
See NAT'L COMM'N ON FED. ELECTION REFORM, supra note 9, at 60-66; BRENNAN CTR.
FOR JUST. AT N.Y.U. SCH. OF LAW, CITIZENS WITHOUT PROOF: A SURVEY OF AMERICANS' POSSESSION OF DOCUMENTARY PROOF OF CITIZENSHIP AND PHOTO IDENTIFICATION 3 (2006),
http://www.brennancenter.org/dynamic/subpages/downloadfile_39242.pdf; JOHN PAWASARAT, THE
DRIVER LICENSE STATUS OF THE VOTING AGE POPULATION IN WISCONSIN 4-5 (2005),
http://www.uwm.edu/Dept/ETlI/barriers/DfiversLicense.pdf.
15.
2005).
Borders v. King County, No. 05-2-00027-3 (Wash. Super. Ct. Chelan County June 24,
Michigan Law Review
[Vol. 105:631
which concerns about fraud would arise-without photo identification, such
as traveling by bus and subway, making credit card purchases via telephone,
accessing pornography over the Internet, and voting via absentee ballot.
More important, erroneous exclusion of legitimate participants carries
greater costs in the voting context because assessing the will of the people
as a whole is an essential objective of democracy.
Politicians and opinion leaders critical of photo-identification proposals
regularly recite talking points about threats to voter participation by the poor
and minorities but often fail to quantify this assertion or elaborate on the
value of widespread participation. Widespread participation furthers democratic legitimacy by producing a government that reflects the will of the
people and
allowing
of citizens to hold government officials
accontale
or teir diverse
. • groups
16
accountable for their decisions. Various constitutional and statutory provisions promote broad participation by eliminating voter qualifications that
many believed were reasonable, such as paying a two-dollar poll tax or exhibiting an ability to read. As the U.S. Supreme Court stated, "Especially
since the right to exercise the franchise in a free and unimpaired manner is
preservative of other basic civil and political rights, any alleged infringement of the right of citizens to vote must be carefully and meticulously
scrutinized."' 7
This Article engages in a careful and meticulous analysis of the conceptual, empirical, and legal issues arising from photo-identification proposals.
In addition, the Article applies an empirical approach that has the potential to reframe various election law controversies. Current scholarship often
rests upon isolated democratic goals and unsubstantiated factual assumptions. Election law, however, involves competing values, such as access and
integrity. Votes provide a metric that allows for costs and benefits to be
quantified. Instead of relying on personal assumptions about how politics
works, 9 scholars and lawmakers should use data to resolve controversies
such as how many fraudulent voters relative to legitimate voters are excluded by photo-identification requirements, partisan challengers at the
polls, restrictions on voter-registration organizations, and various methods
of purging voting rolls.
This approach also helps in balancing access and fiscal restraint. For example, if voting lines during presidential elections average an hour, how
much would it cost to reduce lines to thirty minutes, fifteen minutes, or five
minutes? What societal gains are realized through increased productivity by
16.
See infra Section III.B.
17.
Reynolds v. Sims, 377 U.S. 533, 562 (1964).
Only a few legal scholars have emphasized empirical data in the law of democracy con-
18.
text. See, e.g., Richard H. Pildes, The Politics of Race, 108 HARV. L. REV. 1359, 1360-62 (1995)
(reviewing QUIET REVOLUTION IN THE SOUTH (Chandler Davidson & Bernard Grofman eds.,
1994)).
19.
Cf Daniel P. Tokaji, The Moneyball Approach to Election Reform, ELECTION LAW @
MORITZ, Oct. 18, 2005, http://moritzlaw.osu.edu/electionlaw/comments/2005/051018.php (last
visited Sept. 28, 2006) (asserting that election law should be based on hard data and rigorous analysis rather than merely anecdotes and intuition).
February 2007]
Voter Identification
those who no longer wait an hour to vote, and increased political participation by those who refuse to wait in longer lines? To what extent does
election-day registration enhance turnout and what are the increased administrative costs and risks of fraud? Real data allows for a more honest and
thoughtful discussion about the structure of democracy, which is especially
useful in light of the self-serving platitudes that incumbent politicians often
bring to the debate. While empirical data does not answer all questions, it is
an essential component in the quest for better rules.
Increasingly, other areas of the law reject urban myths and turn to empirical data for insight. The study of law and economics quantifies problems
and analyzes whether the benefits of legal solutions justify their costs. 20 One
prominent scholar observes that "people often deal poorly with the topic of
risk" and asserts that "sensible policymakers should generally follow science and evidence.'
Better data is also essential to determining whether election regulations
pass constitutional and statutory muster. Judges wander into the political
thicket blindly, for example, when they make decisions based on their own
assumptions about fraud and voter access to photo identification rather than
empirical evidence. The extent to which a regulation inhibits legitimate
votes, rather than deters fraudulent ones, reflects the law's overinclusiveness
and its burden on the fundamental right to vote. Better data will also show
whether a photo-identification requirement abridges the franchise contrary
to the Voting Rights Act and the Constitution's prohibition on poll taxes.
Part I of this Article examines the various methods states currently use
to identify voters and the emerging conflict over photo identification as an
absolute requirement to vote. Part II reveals that anecdotes used to justify
photo-identification requirements are often unrepresentative, misleading,
and even false, and shows how oversimplified analogies fall short under
scrutiny. Part III compiles the best existing data on the pervasiveness of
fraud and the number of voters who lack photo identification, and it provides a roadmap for obtaining even better empirical information. Part IV
explains how data plays a critical role in assessing the constitutional and
statutory status of photo-identification requirements, and Part V reviews
several less restrictive alternatives to photo-identification requirements.
20.
See, e.g., STEPHEN BREYER, REGULATION AND ITS REFORM (1982); ROBERT COOTER &
THOMAS ULEN, LAW AND ECONOMICS (1988); RICHARD A. POSNER, ECONOMIC ANALYSIS OF LAW
(5th ed. 1998); R. H. Coase, The Problem of Social Cost, 3 J.L. & ECON. 1 (1960).
21.
Cass R. Sunstein, The Laws of Fear, 115 HARV. L. REV. 1119, 1123 (2002) (reviewing
PAUL SLOVIC, THE PERCEPTION OF RISK (2000)); cf STEPHEN BREYER, BREAKING THE VICIOUS
CIRCLE: TOWARD EFFECTIVE RISK REGULATION 10-29, 59-68 (1993) (asserting that regulators
devote extensive resources to insignificant problems and too few resources to significant problems,
and proposing risk specialists to assess risk and redirect regulatory resources); Wendy E. Wagner,
Commons Ignorance: The Failure of Environmental Law to Produce Needed Information on Health
and the Environment, 53 DUKE L.J. 1619 (2004) (commenting on the need for better data in formulating environmental law).
Michigan Law Review
I. THE
[Vol. 105:631
VOTER IDENTIFICATION LANDSCAPE
The U.S. Supreme Court's decision in Bush v. Gore ratified presidential
election returns that George W. Bush received roughly one more vote than
Al Gore for every 11,100 votes cast in Florida, and reminded the nation that
every vote counts in a closely divided political environment.
In response, civil rights activists focused largely on reforms designed to
improve access, such as replacing obsolete punch card machines that had
relatively high voter-error rates, providing provisional ballots to voters
whose names did not appear on the voting rolls, and restoring voting rights
to felons who had completed their prison sentences.
An alternative movement characterized fraud as the most significant
threat to democracy. Political groups that purported to assist senior citizens
with voting effectively cast absentee ballots for those with dementia.2 Poll
workers stuffed ballot boxes to benefit their favored candidate. 23 Ineligible
voters, such as former felons, noncitizens, nonresidents, and people who had
already voted, cast illegal ballots with impunity. The National Voter Registration Act of 1993 had worsened these problems, advocates argued, because
it had limited the extent to which officials could purge deadwood from the
registration rolls.2 According to integrity advocates, a photo-identification
2
requirement at the polls would solve some of these problems.
The claims about voter fraud arose from an earlier movement that focused on the integrity of elections in the 1960s. Democrat John F. Kennedy
beat Republican Richard Nixon by only 0.2 percent of the popular vote in
the 1960 presidential contest, and some alleged that fraud in Texas and Illinois cost Nixon the election.27 Republicans responded by organizing
"Operation Eagle Eye," an intricate antifraud campaign designed to detect
22. JOHN FUND, STEALING ELECTIONS: How VOTER FRAUD THREATENS OUR DEMOCRACY
44, 47 (2004); see also Jason H. Karlawish et al., Addressing the Ethical, Legal, and Social Issues
Raised by Voting by Persons with Dementia, 292 J. AM. MED. ASS'N. 1345, 1348 (2004).
23.
See FUND, supra note 22, at 8.
24. Id. at 4, 23-25, 41-55. The National Voter Registration Act, otherwise known as the
"Motor Voter" law, directs states to make "a reasonable effort to remove the names of ineligible
voters from the official lists of eligible voters" when voters have died or moved to another jurisdiction, but also prevents states from removing voters for failing to vote unless they have not voted in
two or more consecutive elections. National Voter Registration Act of 1993, 42 U.S.C. § 1973(a)(4)
(2000).
25.
See FUND, supra note 22, at 136-39.
26. Although, following the 1960 presidential election, the antifraud movement took on a
national partisan cast that implicated race, concerns about fraud and voter suppression existed decades earlier. In 1927, the Committee on Election Administration of the National Municipal League
called for "improving the registration machinery for the purpose of preventing fraudulent voting."
EARL R. SIKES, STATE AND FEDERAL CORRUPT-PRACTICEs LEGISLATION 58-60 (1928). The Com-
mittee asserted that "the present registration systems [did] not properly provide for the purging of
dead wood from the registration lists." Id. at 59. In response, thirty-eight states passed statutes to
deal with the problem. Id. at 70; see also ANDREW GUMBEL, STEAL THIS VOTE: DIRTY ELECTIONS
AND THE ROTTEN HISTORY OF DEMOCRACY IN AMERICA 14-15 (2005) (describing rampant voter
fraud in nineteenth-century America). Additionally, several state political parties employed ballot
protection teams to challenge voters' literacy and citizenship at the polls prior to 1960.
27.
GUMBEL, supra note 26, at 161-67.
Voter Identification
February 2007]
and eliminate unqualified voters from registration rolls and challenge the
qualifications of suspicious voters at the polls. Political operatives also
hoped that Operation Eagle Eye would deter fraud through press coverage
of the security program and photography of voters at polling places.2 ' Republicans deployed tens of thousands of poll challengers in the 1964
presidential election, many of whom were concentrated in thirty-six major
metropolitan Democratic strongholds.29 Democrats and civil rights groups
charged that Operation Eagle Eye deterred legitimate voter participation and
intimidated voters of color.3° Similar ballot security efforts continued in subsequent elections, accompanied by claims of voter suppression."
Following the closely contested 2000 presidential contest, Congress
passed the Help America Vote Act of 2002.32 The Act was a broad election
reform package that reflected a series of compromises between Democrats
largely interested in access and Republicans focused on fraud prevention.
The Act enhanced access by providing provisional ballots to registered voters whose names did not appear on the rolls,33 but the law also required all
first-time voters who registered by mail to provide photo or nonphoto
documentary identification (such as a utility bill or bank statement) when
they arrived at the polls.
identify themselves.
34
States remain split as to how other voters must
A. Existing State Laws for Identifying Voters
As of 2006, only Georgia, Indiana, and Missouri required photo identifi-
cation as an absolute condition to vote.35 The other forty-seven states fell
into four general categories. 36
28.
DAVIDSON ET AL., supra note
29.
Id. at 26.
30.
Id. at 35.
13, at 25-31.
31. See id. at 40-95 (documenting ballot security programs from 1968 to 2004, and detailing
thirteen case studies of "ballot security excesses").
32. As discussed above, Congress adopted many of the recommendations proposed by the
2001 Carter-Ford Commission on Federal Election Reform.
33.
42 U.S.C. § 15482(a) (Supp. III 2003).
34. Id. § 15483(b). The Help America Vote Act requires that voters produce a copy of "valid
photo identification or ... a copy of a current utility bill, bank statement, government check, paycheck, or other government document that shows the name and address of the voter." Id.
35. GA. CODE ANN. § 21-2-417 (Supp. 2006), enjoined by Lake v. Perdue, No. 2006-CV119207 (Ga. Super. Ct. Sept. 19, 2006), http://moritzlaw.osu.edu/electionlawlitigation/documents/
Statelnjunction.pdf, and Common Cause/Ga. v. Billups, No. 4:05-CV-0201-HLM (N.D. Ga. Sept.
15, 2006), http://moritzlaw.osu.edu/electionlaw/itigation/documents/georgia9-15-06ordergranting
preliminaryinjunction.pdf;
IND. CODE ANN.
§ 3-11-8-25.1 (West Supp. 2006); Mo.
ANN. STAT.
§ 115.427 (West Supp. 2006), invalidated by Weinschenk v. State, No. SC 88039, 2006 WL
2959284 (Mo. Oct. 16, 2006) (en banc). Florida requests that voters provide a photo identification
card, but unlike Georgia, Indiana, and Missouri, offers voters a provisional ballot that will be
counted if the voter's signature on the provisional ballot matches the voter's signature on the registration list. See infra note 44.
36. While states outside of Georgia, Indiana, and Missouri generally fall into one of the four
categories listed below, some states provide additional detailed rules. Alaska, for example, allows a
Michigan Law Review
[Vol. 105:631
No documentary identification required. In 2005, two-thirds of the U.S.
population lived in the majority of states that did not request documentary
evidence at the polls beyond federal requirements for first-time voters.37 In
these states, poll workers check a voter's name off of preprinted lists of registered voters when he or she arrives at the polls to cast a vote. Voters
establish their identity through various methods, such as signing an affidavit
under penalty of perjury, stating their name, taking an oral oath,39 reciting
their birth date and address to the poll worker, 40 or signing a pollbook that is
compared to the signature on the voter's file.
Documentary identification requested, not required.A handful of states
request that voters produce documentary identification and give them the
option to produce either a photo-identification card, such as a driver's license, or a nonphotographic form of identification, such as a utility bill,
bank statement, government check, or paycheck. 42 In these states, voters
voter who lacks documentary identification to cast a ballot if he or she is identified by poll workers.
ALASKA STAT. § 15.15.225 (2004). Voters in Louisiana who lack photo identification are subject to
challenge. LA. REV. STAT. ANN. §§ 18:562, 18:565 (2004 & Supp. 2006). Many states that require
documentary identification as an absolute requirement to vote allow those without documentary
identification to cast a provisional ballot that officials will count if the voter presents the proper
documentation to an appropriate election official within one or two days. See, e.g., COLO. REV.
STAT. § 1-7-110 (2005).
37. As of July 1, 2005, the combined population of the District of Columbia and the twentyeight states that did not request documentary identification at the polls was 196,194,611 out of a
total U.S. population of 296,410,404. U.S. CENSUS BUREAU, ANNUAL ESTIMATES OF THE POPULATION FOR THE UNITED STATES AND STATES, AND FOR PUERTO Rico: APRIL 1, 2000 TO JULY 1, 2005
(2005), http://www.census.gov/popestlstates/NST-ann-est.html. The full list of such states include
California, the District of Columbia, Idaho, Illinois, Iowa, Kansas, Maine, Maryland, Massachusetts,
Michigan, Minnesota, Mississippi, Nebraska, Nevada, New Hampshire, New Jersey, New York,
North Carolina, Ohio, Oklahoma, Oregon, Pennsylvania, Rhode Island, Texas, Utah, Vermont, West
Virginia, Wisconsin, and Wyoming. Although Kansas and Pennsylvania do not request documentary
evidence from most voters, they require all first-time voters-not just first-time voters who registered by mail, as required of all states by the federal Help America Vote Act-to produce
documentary identification at their polling place to cast a vote. KAN. STAT. ANN. § 25-2908 (Supp.
2005); 25 PA. STAT. ANN. § 3050 (West Supp. 2006). In 1996, Michigan passed a law that requested
state-issued photo identification but allowed voters without identification to sign an affidavit to
establish their identity. MICH. Comp. LAWS ANN. § 168.523 (West Supp. 2006). The law was never
implemented, however, because the Michigan Attorney General issued an advisory opinion that
found the identification requirement unconstitutional. Dawson Bell, Court Jumps Into Dispute
over Voter ID Checks, DETROIT FREE PRESS, Apr. 27, 2006, at News 1, available at
http://www.freep.conapps/pbcs.dll/article?AD=/20060427/NEWS06/604270623&template=printa
rt. At the request of Republican state legislators, the Michigan Supreme Court recently agreed to
issue an advisory opinion on the constitutionality of the 1996 law. (The five Republican-nominated
justices voted to issue the advisory opinion and the two Democratic nominees opposed issuing the
opinion.) Id.
38.
E.g., IOWA CODE ANN. § 49.77 (West 1999).
39.
E.g., CAL. ELEC. CODE § 14243 (West 2003).
E.g., MD. CODE ANN., ELEC. LAW § 10-310 (LexisNexis Supp. 2005).
41. E.g., N.J. STAT. ANN. § 19:15-17 (West Supp. 2006). Although Oregon now conducts its
elections by mail, OR. REV. STAT. § 254.465 (2005), county clerks are nonetheless required to maintain some physical polls, OR. REV. STAT. § 254.474 (2005), and voters who opt to cast a ballot in
person establish their identity by signing a poll book, OR. REV. STAT. § 254.385 (2005).
40.
42. In 2005, states that requested documentary identification but provided an affidavit option
or other means for those without documentary identification to vote included Arkansas, Connecticut,
Delaware, Kentucky, North Dakota, and Tennessee.
Voter Identification
February 20071
who do not bring documentary identification to the polls can establish their
identity by affidavit or other means.
43
Photo identification requested, not required.A few states request that a
voter produce a form of photo identification but provide other avenues for a
voter who lacks photo identification to establish her identity, such as by
signing an affidavit or reciting her birth date and address."
Documentary identification required. Just under a dozen states require
documentary identification as an absolute requirement to vote.45 Acceptable
identification generally includes photo identification, or nonphoto identifica46
tion such as a utility bill or bank statement. Thus, these states effectively
expand the Help America Vote Act's documentary requirements for firsttime voters who registered by mail to all voters.
B. Photo-IdentificationRequirements to Vote
4
In 2005, Republican-controlled legislatures in Georgia and Indiana 1
passed laws mandating government-issued photo identification as an absolute requirement to vote at the polls. 49 Georgia's new statute reduced the
43. In North Dakota, a voter without photo identification can vote without being challenged
by providing her date of birth-provided that a member of the election board or a clerk knows her
personally, and will vouch that the voter is qualified. N.D. CENT. CODE § 16.1-05-07 (Supp. 2005).
A voter who is not recognized by poll workers may still vote if he signs an affidavit stating that he is
a qualified voter. N.D. CENT. CODE § 16.1-05-06 (2004). In Arkansas, if a voter does not present
documentary identification when asked, the poll worker simply makes a note on the precinct voterregistration list that the voter lacked identification; however, after each election the county board of
commissioners "may review the precinct voter registration lists and may provide the information of
the voters not providing identification at the polls to the prosecuting attorney" who then "may investigate possible voter fraud." ARK. CoDE ANN. § 7-5-305 (Supp. 2005).
44. In 2005, a voter without photo identification could establish his identity by signing an
affidavit in Louisiana and South Dakota, and by reciting his birth date and address in Hawaii. Florida requests that voters produce photo identification, but those without photo identification can cast
a provisional ballot that will be counted if the voter's signature on the Provisional Ballot Voter's
Certificate and Affirmation matches the signature on the voter's registration. FLA. STAT. § 101.048
(2006).
45. As of 2006, these states included Alabama, Alaska, Arizona, Colorado, Montana, New
Mexico, South Carolina, Virginia, and Washington. Arizona is unique in that a voter without photo
identification must produce two pieces of nonphoto documentary identification that have both the
voter's name and address. ARiz. REV. STAT. ANN. § 16-579 (Supp. 2005); cf Purcell v. Gonzalez,
127 S. Ct. 5, 8 (2006) (per curiam) (vacating the Court of Appeals's injunction suspending Arizona's
voter identification rules because of the failure of that court to explain its ruling, and expressing "no
opinion ... on the correct disposition ... or on the ultimate resolution" of the case).
46. A couple of states, however, are more restrictive in the nonphoto documentary identification they require. Virginia, for example, accepts only a voter-registration card, driver's license, any
other identification card issued by Virginia or the federal government, or a photo identification issued by an employer. VA. CODE ANN. § 24.2-643 (Supp. 2005).
47. GA. CODE ANN. § 21-2-417 (Supp. 2006).
48.
IND. CODE
ANN.
§ 3-11-8-25.1 (West Supp. 2006).
49. In the Georgia Senate, thirty-one Republicans voted for the measure, while eighteen
Democrats and two Republicans voted against it. See Georgia General Assembly, Senate Vote 565,
http://www.legis.state.ga.us/legis/2005_06/votes/svO565.htm (last visited Sept. 16, 2006). In the
Georgia House, ninety Republicans and one Democrat voted for it, while seventy-two Democrats
and three Republicans voted against it. See Georgia General Assembly, House Vote 510,
642
Michigan Law Review
[Vol. 105:631
acceptable forms of identification from seventeen-which included nonphoto identification such as bank statements and paychecks-to six forms of
government-issued photo identification. The new law also made photo
identification an absolute requirement to vote at the polls by eliminating an
earlier provision that had allowed a voter without identification to sign an
affidavit. The new Georgia law did not, however, require that an absentee
voter establish his identity through photo identification.5 '
The American Civil Liberties Union, Common Cause, and other groups
brought suit challenging the law under the Voting Rights Act, the Fourteenth
and Fifteenth Amendments to the U.S. Constitution, and other legal provi-
sions. 5' A federal district court granted a preliminary injunction preventing
implementation of the new law, concluding that the law would likely
constitute an undue burden on the right to vote and that fees for photoidentification cards would likely constitute a poll tax.53
The Indiana photo-identification law, which took effect on January 1,
2006, requires that voters provide a photo-identification card issued by the
Indiana state or the federal government.4 The statute includes exceptions for
the "indigent... [who are] unable to obtain proof of identification without
the payment of a fee" and voters whose religious beliefs prevent them from
55
being photographed. A voter who falls into either of those categories may
cast a provisional ballot at the polling place, which will be counted within
http://www.legis.state.ga.us/egis/2005_06/votes/hvO5lO.htm (last visited Sept. 16, 2006); see also
Georgia General Assembly, HB 244, http://www.legis.state.ga.usegis/2005-06/sum/hb244.htm
(last visited Sept. 16, 2006) (legislative history of Senate Bill 84). In the Indiana House, all fifty-two
Republicans who were present voted for the bill; all forty-five Democrats who were present voted
against it. See Indiana General Assembly, Action List: S.B. 0483, http://www.in.gov/apps/lsa/
session/billwatch/billinfoyear=2005&request=getActions&doctype=SB&docno=0483 (last visited
Sept. 16, 2006); Indiana General Assembly, Roll Call 259: Passed, http://www.in.govllegislative/
bills/2005/PDF/Hrollcal/0259.PDF.pdf (last visited Sept. 16, 2006); Mary Beth Schneider, House
OKs Strict Voter ID Bill, INDIANAPOLIS STAR, Mar. 22, 2005, at lB. Similarly, the 33-17 vote in the
Indiana Senate was a straight party vote. Mary Beth Schneider, Voter ID Law Looming for Hoosiers,
INDIANAPOLIS STAR, Apr. 13, 2005, at IA. Republican governors signed the photo-identification
requirement into law in both Georgia and Indiana. Republican-controlled legislatures in five other
states have passed photo-identification laws that Democratic governors subsequently vetoed. FUND,
supra note 22, at 138.
50.
GA. CODE ANN. §§ 21-2-220, 21-2-417 (Supp. 2006); Sonji Jacob & Carlos Campos,
Perdue Signs ID Bill, ATLANTA J.-CoNsT., Apr. 23, 2005, at lB.
51.
See GA. CODE ANN. § 21-2-381 (Supp. 2006) (indicating that any applicant for an absentee ballot must provide her address and identify the primary, election, or runoff in which she intends
to vote).
52. Complaint for Declaratory and Injunctive Relief at 32-41, Common Cause/Ga. v.
Billups, 406 F. Supp. 2d 1326 (N.D. Ga. 2005) (No, 4:05-CV-0201-HLM).
53. Common Cause/Ga., 406 F. Supp. 2d at 1376-77. The Department of Justice refused to
object to Georgia's new photo-identification requirement under Section 5 of the Voting Rights Act,
but the Washington Post later reported that four of five career attorneys recommended objection but
were overruled by Republican political appointees. Dan Eggen, Official's Article on Voting Law
Spurs Outcry, WASH. POST, Apr. 13, 2006, at A19.
54.
IND. CODE ANN.
§ 3-11-8-25.1 (West Supp. 2006).
55. Id. § 3-11.7-5-2.5(c) (West 2006). Voters who live and cast their ballots in a statelicensed healthcare facility are not required to show photo identification. Id. § 3-11-8-25.1(f) (West
Supp. 2006).
February 2007]
Voter Identification
two weeks of the election, but only if the voter makes a separate trip to the
county elections board and signs an indigency or religious-objector affidavit
(such affidavits are not made available to voters at polling places). 6 Like the
Georgia law, the Indiana photo-identification requirement does not require
that an absentee voter establish her identity through photo identification.57
The Indiana Democratic Party filed suit, and the Federal District Court
for the Southern District of Indiana refused to enjoin the law, asserting that
the plaintiffs failed to prove that the photo-identification requirement would
burden voting in violation of the federal Constitution or the Voting Rights
Act."
The photo-identification law adopted in Missouri required that voters
present an unexpired state- or federal-issued photo identification that contains the voter's name as it appears on the voter registration rolls.5 9 The
statute did not provide free photo-identification cards to indigent citizens.
The Supreme Court of Missouri struck down the provision, reasoning that it
violated the Missouri Constitution's equal protection clause and right to vote
provision.
In September 2005, the Carter-Baker Commission recommended that
the remaining states adopt a photo-identification requirement. 6, The Commission connected its photo-identification proposal to the "Real ID" Act,
which prohibits states from issuing a driver's license or nondriver's identification card after 2007 unless an individual presents documentary proof of
her full legal name and date of birth, Social Security number, and citizenship. 62 The Carter-Baker Commission recommended 63that states require a
"Real ID" card as a prerequisite for voting at the polls.
To mitigate access concerns, the Commission proposed that states "undertake their best efforts to make voter registration and ID accessible and
available to all eligible citizens" through mobile offices and offering "Real
ID" cards to nondrivers free of charge. 64 Further, the Commission recommended that through 2009 states permit voters without a "Real ID" card to
6
cast a provisional ballot by signing an affidavit attesting to their identity. 1
56. Id. ("[AIII provisional ballots must be counted by not later than noon on the second
Monday following the election.").
57.
Id. § 3-11-10-1.2.
58. Ind. Democratic Party v. Rokita, No. l:05-CV-0634-SEB-VSS, 2006 WL 1005037, at
*35, *47 (S.D. Ind. Apr. 14, 2006).
59.
60.
2006).
61.
IND. CODE ANN.
§ 3-11-8-25.1
(West Supp. 2006).
Weinschenk v. Missouri, No. SC88039, 2006 Mo. LEXIS 122, at *30 (Mo. Nov. 7,
COMM'N ON FED. ELECTION REFORM, supra note 4, at
62.
Id. at 19-21.
63.
Id. at21.
64.
Id. at 21, 33-34.
18-21.
65. Id. at 21. Former President Jimmy Carter asserted that the proposal's transition period
and card distribution proposals mitigate access problems, and he criticized photo-identification
legislation that failed to incorporate these elements. See Letter from Jimmy Carter, Former U.S.
Michigan Law Review
[Vol.
105:631
[
Thereafter, citizens would be allowed to vote without a "Real ID," but their
ballots would only be counted if they returned with valid photo identification within forty-eight hours-thereby making the showing of a "Real ID"
card an absolute requirement to vote. 6 The Commission also proposed that
states confirm the identity of absentee voters not through "Real ID," but
61
through signature matching.
Before states follow the lead of the Georgia, Indiana, and Missouri state
legislatures and the Carter-Baker Commission, however, lawmakers should
pause to examine closely the arguments put forth in support of photoidentification requirements.
II.
THE SHORTCOMINGS OF ANECDOTE, ANALOGY, AND INTUITION
TO JUSTIFY PHOTO IDENTIFICATION
Photo-identification advocates often rely on two categories of assertions:
(1) anecdotes about voter fraud, and (2) analogies to other contexts that require photo identification. Both are deeply flawed.
Voter-fraud anecdotes are often misleading, incomplete, and unrepresen-
tative. Advocates selectively emphasize the anecdotes that are sure to evoke
indignation or other emotions rather than the most typical fraud incidents
and omit facts or other stories that cut against their desired policy result.
They also employ analogy to justify their proposals, but they often ignore
important differences between voting and activities that require photo identification, such as traveling by air and purchasing alcohol.
A. Misleadingand UnrepresentativeAnecdotes about Voter Fraud
Voter-fraud anecdotes can lead to misleading generalizations absent disclosure of the anecdotes' truthfulness and typicality. We cannot determine
whether a photo-identification requirement is an appropriate response to
voter fraud, for example, unless we understand whether there are ten fraudulent votes for every one hundred, ten thousand, or one million votes cast.
68
Professor Michael Saks explains as follows:
[A]necdotal evidence is heavily discounted in most fields, and for a perfectly good reason: such evidence permits only the loosest and weakest of
inferences about matters a field is trying to understand. Anecdotes do not
President, to Robin Carnahan, Missouri Sec'y of State (Mar. 16, 2006), available at
http://www.sos.mo.gov/img/03-16-06-PresidentCarterLetter.pdf (asserting that some Missouri
legislators improperly invoked the Carter-Baker Commission photo-identification proposal to support Missouri photo-identification legislation because the Missouri bill did not contain adequate
safeguards for voter access).
66.
COMM'N ON FED. ELECTION REFORM, supra note 4, at 21.
67.
Id. at 20.
68. "The significance of a story of oppression depends on its representativeness .... [T]o
evaluate policies for dealing with the ugliness we must know its frequency, a question that is in the
domain of social science rather than of narrative." Richard A. Posner, Legal Narratology,64 U. CHI.
L. REv. 737, 742 (1997).
February 2007]
Voter Identification
permit one to determine either the frequency of occurrence of something or
its causes and effects ....
These anecdotes may work as a persuasive device, in that a few examples of apparent greed, abuse, or system irrationality can arouse people
emotionally....
[Some anecdotes] are systematically distorted portrayals of the actual
cases they claim to report .... Even when true, anecdotes enjoy a persuasive power that far exceeds their evidentiary value.
Anecdotes have a power to mislead us into thinking we know things
that anecdotes simply cannot teach us. 69
Professor David Hyman illustrates the shortcomings of anecdote in pol70
icy-making by recounting a story conveyed by President Ronald Reagan.
For years Reagan told the story of an alleged "welfare queen" who he
claimed used eighty different names and a dozen Social Security cards to
defraud the government of more than $150,000. Even after the true story
was pointed out to him-the woman had used two aliases to take $8,000Reagan continued to use his false version.7' The reliance on anecdote to discredit the welfare system became common. One white waitress in suburban
Chicago complained that "'blacks buy porterhouse steaks with food stamps,
while we eat hamburgers.' ,72 The woman admitted that she "had never actually seen any blacks do this. But she [had] heard and read stories, and that
[was] enough. 73
Anecdotes about voter fraud are also misleading and fail to indicate the
frequency of the alleged fraud.74 For example, although John Kerry lost the
2004 presidential race nationwide, he won Wisconsin by just eleven thousand votes. Republicans suspected that massive fraud swung the Wisconsin
election to Kerry and pushed -for a photo-identification requirement at the
polls.
In August 2005, Republican politicians in Wisconsin held a press conference to emphasize the need for a photo-identification requirement. The
Republicans announced that their research had uncovered nine people who
in November 2004 had voted in Milwaukee and also cast ballots in Chicago,
69.
Michael J. Saks, Do We Really Know Anything About the Behavior of the Tort Litigation
System-and Why Not?, 140 U. PA. L. REv. 1147, 1159-61 (1992).
70.
See David A. Hyman, Lies, DamnedLies, and Narrative,73
71.
See The Mendacity Index, WASH.
MONTHLY,
IND.
L. J. 797, 804 (1998).
Sept. 1,2003, at 27.
72. Hyman, supra note 70, at 804 n.28 (quoting Isabel Wilkerson, The Tallest Fence: Feelings on Race in a White Neighborhood, N.Y. TIMES, June 21, 1992, § 1, at 18).
73.
Id.
74. My utilization of an anecdote of misleading anecdotal evidence in Wisconsin should not
be construed to suggest that all anecdotes about fraud are misleading, false, or otherwise flawed.
Instead, the Wisconsin anecdote illustrates the flaws of anecdote and the need for empirical data to
determine the frequency and typicality of voter fraud.
Michigan Law Review
[Vol. 105:631
Minneapolis, or Madison. 5 The press conference was held in front of one of
the addresses allegedly used to vote twice, according to Wisconsin GOP
Chair Rick Graber.16 "'We now are able to make this link to show that this
voter fraud has crossed state lines,'" announced Republican State Representative Jeff Stone."
In its September 2005 Report, the Carter-Baker Commission also supported its call for photo identification by invoking the 2004 Wisconsin
election:
In Milwaukee, Wisconsin, investigators said they found ... more than 200
cases of felons voting illegally and more than 100 people who voted twice,
used fake names or false addresses, or voted in the name of• a 781
dead person.
Moreover, there were 4,500 more votes cast than voters listed.
Commissioner Susan Molinari, a Republican and former congresswoman, asserted that a photo-identification requirement was justified
because the election in Wisconsin was "decided by illegal votes," a fact "established by a joint report written by the U.S. Attorney, FBI, Chief of Police
and senior local election official-both Republicans and Democrats."79
But these Wisconsin anecdotes are misleading. Of the nine "doublevoting" individuals named by the Republican Party leadership at their press
conference, none have been indicted for fraud by the Republican-appointed
U.S. attorney.80 Six of the cases involved clerical errors, and in three cases
individuals with a similar name but a different birth date voted in Chicago,
Madison, or Minneapolis."
In its support for a photo-identification requirement, the Carter-Baker
Commission also failed to disclose a variety of important factors regarding
the Wisconsin anecdote. First, the ballots examined by the Joint Task Force
Investigating Possible Election Fraud differ from those in other states. Most
states require voters to register in advance of election day and restrict the
casting of regular ballots to those on the voting rolls.8' Wisconsin and a few
other states, however, have election-day registration, and thus unregistered
75.
Greg J. Borowski, Nine May Have Voted in 2 Cities,
MILWAUKEE J. SENTINEL, Aug.
10,
2005, at B 1.
76. Voter ID Gets Push from GOP: Milwaukee Cases Cited as Example, CAP.
son, Wis.), Aug. 10, 2005, at 3A.
77.
TIMES
(Madi-
Id.
COMM'N ON FED. ELECTION REFORM, supra note 4, at 4 (citing JOINT TASK FORCE INVESTIGATING POSSIBLE ELECTION FRAUD, PRELIMINARY FINDINGS OF JOINT TASK FORCE
78.
(2005), http://www.wispolitics.com/1006/electionfraud.
pdf); see also id. at 18 n. 19 (establishing that fraud and multiple voting occur by referring back to
Section 1.
1, which details alleged fraud in November 2004 elections in Washington State and Wisconsin).
INVESTIGATING POSSIBLE ELECTION FRAUD
79.
Id. at 90 (additional statement of Commissioner Susan Molinari).
80.
See Borowski, supra note 75.
81.
Id.
82. Under the Help America Vote Act, an individual who is not on the voting rolls may cast a
provisional ballot, which is counted if officials later determine that the individual is a properly registered voter. 42 U.S.C. § 15482(a) (Supp. 1112003).
February 2007]
Voter Identification
individuals can show up, register, and cast a vote. 3 The Task Force investigation focused on 70,000 votes of individuals who registered at the polls on
election day in Milwaukee, a pool of votes that did not exist in forty-four
other states.
Further, many of the fraudulent activities listed by the Carter-Baker
Commission are unrelated to photo identification. A photo-identification
requirement would not have kept ineligible felons from voting, nor would ''it
have prevented the final total of "4,500 more votes cast than voters listed. 8
Out of the 70,000 same-day registrations studied, investigators found just
over one hundred questionable instances in which people may have voted
twice, used
false addresses or fake names, or voted in the name of a dead
85
person.
Assuming that each of these instances resulted from intentional voter
fraud rather than a clerical mistake or other reason, this is a fraud rate of less
than one-seventh of one percent (0.14% to be exact), or one in seven hundred. And the rate may not be that high. As of December 2005, authorities
had charged only four people out of the group with double-voting,
and three
6
of the charges resulted in dismissal, acquittal, and a hung jury.
Contrary to the claims of Carter-Baker Commissioner Molinari, the Task
Force did not find that the Wisconsin election was "decided by illegal
votes."' Even in the improbable event that all one hundred alleged fraudulent votes and two hundred improper felon votes were cast for John Kerry,
his lead in the state would have been reduced from 11,000 to 10,700. The
U.S. attorney explicitly stated, "We don't see a massive conspiracy to alter
the election in Milwaukee, one way or another." 8
83.
Idaho, Maine, Minnesota, Montana, New Hampshire, and Wyoming also allow voters to
register on election day.
supra note 4, at 4.
84.
COMM'N ON FED. ELECTION REFORM,
85.
JOINT TASK FORCE INVESTIGATING POSSIBLE ELECTION FRAUD,
supra note 78, at 2.
86. Steve Schultze, No Vote Fraud Plot Found: Inquiry Leads to Isolated Cases, Biskupic
Says, MILWAUKEE J. SENTINEL, Dec. 6,2005, at Al.
87. COMM'N ON FED. ELECTION
missioner Susan Molinari).
REFORM,
supra note 4, at 90 (additional statement of Com-
88. Schultze, supra note 86, at Al (quoting U.S. Attomey Steve Biskupic). The Commission
on Federal Election Reform also cited ex-felon voting and votes cast in the names of the dead as
evidence of fraud in a closely contested 2004 Washington gubernatorial race that was decided by
129 votes. COMM'N ON FED. ELECTION REFORM, supra note 4, at 4. In a separate statement, Commissioner Susan Molinari argued that states should adopt photo-identification requirements because
the Washington race was "decided by illegal votes" and "this fact was established by a lengthy trial
and decision of the court." COMM'N ON FED. ELECTION REFORM, supra note 4, at 90 (additional
statement of Commissioner Molinari). These claims suffer from many of the problems of the Wisconsin anecdote. A photo-identification requirement would not have stopped ex-felon voting in
Washington State. The Commission also failed to note that the Washington court had concluded that
of more than 2.8 million ballots, only six had been cast by voters who had voted twice and only
nineteen had been cast in the name of deceased individuals. Borders v. King County, No. 05-200027-3 (Wash. Super. Ct. Chelan County June 24, 2005). Since the margin of victory was 129
votes, it is clear that these twenty-five illegal votes (many of which would not have been prevented
by a photo-identification requirement) did not decide the election, even making the improbable
assumption that all of them went to Democratic candidate Christine Gregoire. Further, the Commission did not emphasize that most, if not all, of the nineteen votes cast statewide in the names of the
Michigan Law Review
[Vol. 105:631
Photo-identification advocates generally respond to these observations
by emphasizing the existence of fraud rather than its magnitude. After the
U.S. attorney in Wisconsin announced that no massive conspiracy of voter
fraud had been found, the GOP released a statement indicating that "the
Republican Party of Wisconsin continues to maintain that one case of voter
fraud is one too many."8 9 The Carter-Baker Commission also dismissed the
need to examine the extent of empirical evidence:
While the Commission is divided on the magnitude of voter fraud-with
some believing the problem is widespread and others believing that it is
minor-there is no doubt that it occurs. The problem, however is not the
magnitude of the fraud. In close or disputed elections, and there are many,
a small amount of fraud could make the margin of difference. 90
The magnitude of fraud, however, is critical to determining whether a
photo-identification requirement will do more harm than good. One cannot
assess a photo-identification requirement's true cost without determining
whether, for every ten cases of voter fraud, a photo-identification requirement would deter from voting one, one hundred, or ten thousand legitimate
voters. Depending on the magnitude of fraud, a photo-identification requirement could erroneously skew election outcomes to a greater extent than
would a lack of such a requirement.
In addition to overlooking typicality, anecdotes often distract with emotion and fail to reveal the causes or effects of fraud. 9' On the first page of his
book Stealing Elections: How Voter FraudThreatens Our Democracy, Wall
Street Journal editor John Fund asks, "How sloppy [is our electoral system]? Lethally so. At least eight of the nineteen hijackers who attacked the
World Trade Center and the Pentagon were actually able to register to vote
in either Virginia or Florida while they made their deadly preparations for
9/1 1. 92
dead were cast absentee, and thus would not have been prevented by a photo-identification requirement for in-person voting (the Commission recommended a signature requirement over photo
identification for absentee voting). See Gregory Roberts, Six More Charged with Offenses in 2004
Election, SEATTLE POST-INTELLIGENCER, June 22, 2005, at B 1.
89. Statement of the Wisconsin Republican Party Chairman, Statement Re: U.S. Attorney
Biskupic's investigation into voter fraud Dec. 6, 2005, available at http://www.wisgop.org/site/
Viewer.aspx?iid=23 1&mname=ArticleGroup&rpid=802 (last visited Feb. 24, 2006).
90.
COMM'N ON FED. ELECTION
91.
Professor Richard Epstein states the following:
REFORM,
supra note 4, at 18 (emphasis added).
The capacity of narrative to inflame, inform, or excite depends on its ability to take you away
from the peak of the distribution to see what some extraordinary novel and different circumstance is and indeed that is exactly why we call these things novel because of the way in which
they take you away from the core. But if you are trying to understand the way in which social
reality works then the important thing to remember is that the prosaic and the boring is often
far more important in the way in which the world organizes itself than is the exotic and pro-
fane.
Discussion, Legal Scholarship and Disciplinary Politics,45 STAN. L. REV. 1671, 1678 (1993) (remarks of Professor Richard Epstein), cited in Hyman, supra note 70, at 836.
92.
FUND, supra note 22, at 1.
February 2007]
Voter Identification
Photo-identification proponents rely on this dramatic statement to cite
the potential for voter fraud. One editorialist, for example, claimed that had
the hijackers "survived, they could have shown up on Election Day and
voted. 93 But it remains unclear that eight of the hijackers were registered to
vote--data has not yet been found to confirm this assertion. 94 Even assum-
ing eight of the hijackers registered to vote in Virginia or Florida, it is
unlikely that the registrations caused the lethal attack on 9/11. Fund does not
reveal how many of the improper registrations resulted in fraudulent votes.
Further, the nineteen hijackers obtained sixty-three driver's licenses from
various states and "could have shown up on Election Day and voted," even if
a photo-identification requirement to vote had been in effect.9
Photo-identification advocates also often cite irregularities that would
not be prevented by a photo-identification requirement. For example, proponents regularly cite the registration of fictitious people, illegal aliens, and
pets, and the fact that voting rolls contain more names than U.S. Census
96
records as a justification for photo-identification requirements. These advocates fail to disclose that many bloated voting rolls are not inflated by
malicious citizens who plan to vote in multiple jurisdictions. Instead,
bloated rolls are often caused by county registrars' failure to purge old data
after voters move. Further, photo-identification advocates do not provide
evidence that most fictitious registrations are caused by people who vote
under their own name, a second time as "Elmer Fudd," and a third time as
93. Editorial, One Lawyer, One Vote, INVESTOR'S Bus. DAILY, Oct. 22, 2004, at A16. Members of Congress used this same argument during the debate over the Help America Vote Act. See,
e.g., 148 CONG. REC. Si 171, 1176 (daily ed. Feb. 26, 2002) (statement of Sen. Bond) ("In Colorado,
a Saudi man detained by Federal authorities for questioning about the September 11 terrorist attacks
voted in Denver during last year's Presidential election, even though he was not a U.S. citizen ...In
North Carolina, a Pakistani man facing a vote fraud charge has been linked to at least two of the
September 11 hijackers.").
94. When my research assistant Daniel Taylor contacted John Fund and asked about the
source of the assertion that eight of the hijackers were registered in either Florida or Virginia, John
Fund indicated that he had obtained the fact from an interview with then-Assistant Attorney General
Michael Chertoff. Taylor then contacted the Department of Justice's Criminal Division, the Counterterrorism Section, and Voting Section, and no one knew about the claim. At the suggestion of these
offices, Taylor filed a FOIA request. He also repeatedly called the Department of Homeland Security (of which Chertoff became the Secretary), but no one responded to Taylor. Taylor also contacted
the former Virginia Secretary of the Board of Elections, Cameron Quinn. Quinn indicated that she
was unable to confirm or deny that the 9/11 hijackers were registered to vote in Virginia. She was
familiar with the claim, and indicated that her office investigated it while she was Secretary of the
Board of Elections. However, they had a difficult time obtaining from federal officials the actual
names of the hijackers, their Social Security numbers (which is how they usually look up registrations), or their actual voter registration numbers. As a result, she believes that her agency was never
able to prove or disprove that any of the 9/11 hijackers registered to vote in Virginia. Taylor's calls
to the Florida Secretary of State were not returned.
95.
See One Lawyer One Vote, supra note 93.
96. See FUND, supra note 22, at 4; cf COMM'N ON FED. ELECTION REFORM, supra note 4, at
4 ("One potential source of election fraud arises from inactive or ineligible voters left on voter registration lists ....[Tihere were over 181,000 dead people listed on the voter rolls in six swing states
in the November 2004 elections .... (citing Geoff Dougherty, Dead Voters on Rolls, Other
Glitches Found in 6 Key States, CHI. TRIB., Dec. 4, 2004, at C13)); id. at 90 (additional statement of
Commissioner Molinari) (asserting that photo-identification requirements are necessary because
"voter rolls are filled with fictional voters like Elmer Fudd and Mary Poppins").
Michigan Law Review
[Vol. 105:631
"Mary Poppins," rather than by workers who get paid two dollars per name
97
registered and profit by padding their registrations with fictitious names.
Such fictitious reporting is a problem, but primarily to the voter-registration
organizations that pay workers for fraudulent names and the jurisdictions
that contend with bloated voter registration rolls. 98
Proponents of photo-identification requirements also regularly rely on
instances of absentee ballot fraud rather than voter fraud at the polls to support their proposals. 99 A photo-identification requirement at the polls,
however, does not prevent absentee ballot fraud. Indeed, in Georgia and
Indiana, absentee voters need not produce photo identification, ° and the
Carter-Baker Commission proposed that states confirm the identity of absentee voters through signature matching rather than photo identification.' O'
The fact that photo-identification advocates use unrepresentative and
misleading anecdotes that would persist even with the implementation of a
photo-identification requirement does not, in and of itself, mean that voter
fraud does not exist. Instead, it simply illustrates the limitations of anecdotal
analysis. Policy-makers need better data about fraud and statistical analysis
to fully understand whether the benefits of a photo-identification requirement justify its costs.
B. FlawedAnalogies
By analogizing voting to other contexts, photo-identification advocates
often avoid the question of whether a photo-identification card will reduce
political participation by legitimate voters. People need photo identification
to board a plane, enter federal buildings, cash a check, use a credit card, rent
a video, and buy cigarettes and alcohol, advocates argue. Why should voting
be an exception to this rule?
Analogy is a common rhetorical tool, but it has limitations. Professor
Cass Sunstein articulated this idea as follows:
Everything is a little bit similar to, or different from, everything else ....
Everything is similar in infiniteways to everything else, and also different
97. Anecdotes, analogies, and intuition-rather than rigorous empirical analysis-lead
judges to make unfounded assumptions. See, e.g., Weinschenk v. Missouri, No. SC88039, 2006 Mo.
LEXIS 122, at *77 (Mo. Nov. 7, 2006) (Limbaugh, J., dissenting) ("Although the majority agrees
that there is some evidence of voter fraud at the voter registration stage, they discount that evidence
as if it had no connection with fraud at the polling place. But why else does voter registration fraud
occur if not to vote persons fraudulently registered?").
98. See Common Cause/Ga. v. Billups, 406 F. Supp. 2d 1326, 1361 (N.D. Ga. 2005) (holding that substantial likelihood exists that Georgia photo-identification requirement is
unconstitutional, and noting that "although Defendants have presented evidence from elections
officials of fraud in the area of voting, all of that evidence addresses fraud in the area of voter registration, rather than in-person voting").
99. See, e.g., id. at 1366; COMM'N ON FED. ELECTION REFORM, supra note 4, at 4;
Deroy Murdock, A Necessary Shaming, NATIONAL REVIEW ONLINE, Sept. 14, 2004, http://
www.nationalreview.com/murdock/murdock200409l40829.asp.
100.
GA. CODE ANN. § 21-2-381 (Supp. 2006); IND. CODE ANN. § 3-11-10-1.2 (West 2006).
101.
COMM'N ON FED. ELECTION REFORM, supra note 4, at 20.
February 2007]
e Voter Identification
from everything else in the same number of ways. At the very least one
needs a set of criteria to engage in analogical reasoning. Otherwise one has
no idea what is analogous to what.
By themselves, factual situations tell us little until we impose some
sort of pattern on them.'0 2
The question in examining photo-identification analogies is whether
democracy sufficiently resembles adult recreation, air travel, and other activities that require photo identification and therefore warrants identical
treatment.
While a photo-identification requirement in voting and other contexts
aims to ensure that a person is who she represents herself to be and/or meets
particular qualifications, the costs of erroneous exclusion differ with voting.
John Fund, for example, asserts that the Clinton administration hypocritically pushed for photo-identification requirements for cigarette purchases
but opposed such requirements for voting.' °3 But for those who consider
widespread participation a critical democratic value, erroneously preventing
a legitimate voter from casting a ballot poses more harm than erroneously
preventing a twenty-two-year-old adult from buying cigarettes.
Erroneous exclusion of air travelers or legitimate credit-card users who
lack photo identification may inconvenience individuals and slow the economy, but these harms are different not only with respect to the type of harm
they prevent, but also in their motivations. In the airline and commercial
contexts, participants do not have "votes" that are weighed relative to one
another to assess the will of the entire citizenry and determine who will
govern society. Liquor stores, airlines, and department stores generally lack
incentives to exclude legitimate consumers, whereas some politicians benefit by reducing turnout among particular demographic populations likely to
vote against them. While the benefits of deterring one terrorist outweigh the
costs of excluding ten thousand "safe" air travelers who lack photo identification, the benefits of excluding one fraudulent voter do not outweigh the
costs of excluding ten thousand legitimate voters. This paradox is all the
more disconcerting because so many of those performing the calculus of
voting regulation stand to gain from erring on the side of implementing
rules that reduce turnout.
A similar cost-benefit analysis explains the lack of photo-identification
requirements in many financial contexts. Merchants lose millions of dollars
a year through credit card fraud, but they generally do not require photo
identification or even a signature when individuals use a credit card at a gas
pump or use credit card numbers online. Empirical data about the extent of
fraudulent transactions and the true costs of a photo-identification requirement help individual merchants determine whether the requirement would
increase or decrease expected profits.
102. Cass R. Sunstein, Commentary, On Analogical Reasoning, 106 HARv. L. REv. 741, 774
(1993) (footnote omitted).
103.
FUND,
supra note 22, at 137.
Michigan Law Review
[Vol. 105:631
Even with nonmonetary objectives, such as terrorism prevention and the
protection of minors, a cost-benefit analysis shapes whether photo identification will be required. For example, despite recent bombings in Israel,
London, and Madrid, the United States still generally does not require
commuters entering a subway or a bus to show photo identification. The
administrative burden of requiring photo identification for all commuters
seems high, while the effectiveness of such a requirement in preventing terrorism seems low. The costs of terrorism on airlines are much higher than
the costs of voter fraud, but officials provide a "safety net" procedure that
allows travelers who fail to bring photo identification to the airport to be
searched and board airplanes. 4 Despite minors' ability to obtain wine, cigarettes, movie rentals, and even free pornographic material via the Internet
without photo identification, lawmakers have not deemed the magnitude of
these problems sufficiently large to outweigh distributors' profits and the
convenience and anonymity provided to adult customers.
Policy-makers also rely on a cost-benefit assessment with regard to political participation. Although absentee ballots pose a greater risk of fraud
than voting at the polls, states generally confirm absentee voter identity
through a signature match rather than requiring that absentee voters show a
photo-identification card to a notary public.' 5 Although foreign nationals
have made political contributions to both Democrats and Republicans in
violation of federal law, the law does not require that every donor produce
photo identification that establishes U.S. citizenship.
A photo-identification requirement could disenfranchise twenty million Americans, and policy-makers should resist the temptation to rush to
adopt the proposal based solely on anecdotes, analogy, and "common
sense" popular assumptions.' 6 Without hard data, many people misperceive risk.'0 7 About four in ten Americans believe, for instance, that flying in
104. See Gilmore v. Gonzales, 435 F.3d 1125, 1129 (9th Cir. 2006) ("The identification policy
requires airline passengers to present identification to airline personnel before boarding or be subjected to a search that is more exacting than the routine search that passengers who present
identification encounter.").
105.
See COMM'N ON FED. ELECTION REFORM, supra note 4, at 20. The Carter-Baker Commission expressly adopted a signature requirement rather than a photo-identification requirement for
absentee ballots. Id.
106. Cf Marc Galanter, Real World Tons: An Antidote to Anecdote, 55 Mo. L. REv. 1093,
1094-95 (1996) (wrapping an argument "in the mantle of common sense [is] certainly cause for
suspicion" (footnote omitted)). Many photo-identification advocates defer to "common sense" rather
than hard data and risk analysis. See, e.g., Editorial, Preserving Election Integrity, LAS VEGAS REV.
J., Jan. 28, 2006, at 14B ("Georgia's [photo identification] bill is a common-sense reform that would
bolster public confidence in the election process."); Patrick Mclheran, Editorial, Election Plot or
Not, Milwaukee's Vote Wasn't Clean, MILWAUKEE J. SENTINEL, Dec. 7, 2005, at A21 (arguing that
Wisconsin's Democratic governor should "compromise on the common-sense safeguard of photo
ID").
107.
See Sunstein, supra note 21, at 1123.
Voter Identifict-tion
February 2007]
an airplane is more dangerous than riding in a car,' °8 even though in reality
people are much more likely to die for every mile they ride in a car than for
every mile they fly in a plane.' A variety of factors skew perception of risk,
including perceived control over a situation, familiarity with a process,
stereotypes, personal fears, outrage, and other emotions.' Data is a critical
component of a reasoned decision-making process.
III. THE
NEED FOR EMPIRICAL EVIDENCE TO BETTER
UNDERSTAND FRAUD AND ACCESS
Before enacting any additional fraud-prevention proposals, including a
photo-identification requirement, it is crucial to understand the scope and
nature of voter fraud. Policy-makers need data on both fraud and access to
the polls to determine whether a photo-identification requirement would
lead to fewer erroneous election outcomes by preventing a large number of
fraudulent votes or result in more erroneous election outcomes by deterring
a larger number of legitimate voters. Empirical information would also indicate whether a photo-identification requirement would disproportionately
exclude groups such as senior citizens, the poor, Americans with disabilities,
and people of color.
To date, no systematic, empirical study of voter fraud has been conducted at either the national or the state level."' This gap in knowledge is
not inevitable. This Part examines the best available data on the fraudulent
votes that a photo-identification requirement would deter and the legitimate
votes it would inhibit. This Part also proposes methods that promise to yield
better data about whether a photo-identification requirement would do more
harm than good.
A. Toward Better Data on the Extent of Fraud
Proponents of photo identification assert that voter fraud exists but is
tough to measure because it is difficult to detect. Even if perfect information
is unobtainable, however, we can secure better data that would allow for
108.
BUREAU OF TRANSP. STATISTICS, U.S. DEP'T OF TRANSP., OMNIBUS SURVEY: HOUSEHOLD SURVEY RESULTS: SUMMARY REPORT: DECEMBER 2000, http://www.bts.gov/programs/
omnibus-surveys/household survey/2000/december/summary-report.html
2006).
(last visited Sept. 16,
109. OFFICE OF HAZARDOUS MATERIALS SAFETY, U.S. DEP'T OF TRANSP., A COMPARISON OF
RISK, http://hazmat.dot.gov/riskmgmt/riskcompare.htm (last visited Mar. 1, 2006).
110. See Ann Bostrom, Risk Perceptions: "Experts" vs. "Lay People ", 8 DUKE ENVTL. L. &
POL'Y F. 101 (1997); Peter M. Sandman, Risk Communication:FacingPublic Outrage, EPA J., Nov.
1987, at 21, 21-22 (1987). Control, familiarity, and emotion need not be removed from all decisionmaking, but in light of the misperception of risk, policy-makers and courts also need real data to
make informed judgments.
111.
DAVIDSON ET AL., supra note 13, at 99; U.S. ELECTION ASSISTANCE COMM'N, supra note
13, at 16-19 (acknowledging that no comprehensive study of voter fraud exists, and recommending
various methods for studying voter fraud, including database comparison, risk analysis, statistical
research, and following up on allegations of fraud in the media and literature).
Michigan Law Review
[Vol. 105:631
reasonable assessments about the amount of voter fraud in U.S. elections.
Three approaches-investigations of voter fraud, random surveys of voters
who purported to vote, and an examination of death rolls-provide a better
understanding of the frequency of fraud. All three approaches have strengths
and weaknesses, and thus the best studies would employ all three to assess
the extent of voter fraud. Further, an accurate estimate of the benefits of a
photo-identification requirement must also consider the amount of fraud that
would persist due to forged photo-identification cards,"' and thus would not
be prevented by a photo-identification requirement.
1. Investigationsand Prosecutions of Voter Fraud
Policy-makers should develop databases that record all investigations,
allegations, charges, trials, convictions, acquittals, and plea bargains regard-
ing voter fraud. Existing studies are incomplete but provide some insight.
For example, a statewide survey of each of Ohio's eighty-eight county
Boards of Elections found only four instances of ineligible persons attempting to vote out of a total of 9,078,728 votes cast in the state's 2002 and 2004
general elections. This is a fraud rate of 0.000044%."' The Carter-Baker
Commission's Report noted that since October 2002, federal officials had
charged eighty-nine individuals with casting multiple votes, providing false
information about their felon status, buying votes, submitting false voter
registration information, or voting improperly as a noncitizen." 4 Examined
in the context of the 196,139,871 ballots cast between October 2002 and
August 2005, this represents a fraud rate of 0.000045% (and note also that
not all of the activities charged would have been prevented by a photo-
identification requirement)."'
A more comprehensive study should distinguish voter fraud that could
be prevented by a photo-identification requirement from other types of
fraud-such as absentee voting and stuffing ballot boxes-and obtain statistics on the factors that led law enforcement to prosecute fraud. The study
would demand significant resources because it would require that research-
112. Cf Donna Leinwand, Tech-Savvy Teens Swamp Police With Fake ID's, USA TODAY, July
2, 2001, at IA, available at http://www.usatoday.comnews/nation/20Ol/07/02/fake-ids.htm ("Computer-savvy teenagers are creating millions of fake driver's licenses despite the holograms and other
high-tech security features that states now put on licenses to thwart forgers. Using the Internet,
anyone willing to break a few laws can be a mass producer of fake IDS ....
); Bush Daughter Used
Fake ID to Buy Alcohol, BIRMINGHAM POST, May 31, 2001, at 11.
113.
COALITION ON HOMELESSNESS AND HOUSING IN OHIO & LEAGUE OF WOMEN VOTERS OF
OHIO, LET THE PEOPLE VOTE: A JOINT REPORT ON ELECTION REFORM ACTIVITIES IN OHIO (2005),
available at http://www.cohhio.org/alerts/Election%20Reform%2OReport.pdf
(finding only four
cases of fraud statewide, based on interviews of the Director or Deputy Director of each of the
state's eighty-eight county Boards of Elections in June 2005, that asked, "Were there any voter fraud
cases within your county from the Election of 2002 and 2004?").
114.
COMM'N ON FED. ELECTION REFORM, supra note 4, at 45 (citing Press Release, U.S.
Dep't. of Justice, Department of Justice to Hold Ballot Access and Voting Integrity Symposium
(Aug. 2, 2005), http://www.usdoj.gov/opa/pr/2005/August/05-ag-404.htm).
115.
See BRENNAN CTR. FOR JUST. & OVERTON, supra note 12, at 9-10.
Voter Identification
February 2007]
ers interview 6 and pore over the records of local district attorneys and election boards.'
Hard data on investigations, allegations, charges, pleas, and prosecutions
is important because it quantifies the amount of fraud officials detect. Even
if prosecutors vigorously pursue voter fraud, however, the number of fraud
cases charged probably does not capture the total amount of voter fraud.
Information on official investigations, charges, and prosecutions should be
supplemented by surveys of voters and a comparison of voting rolls to death
rolls.
2. Random Surveys of Voters
Random surveys could give insight about the percentage of votes cast
fraudulently. For example, political scientists could contact a statistically
representative sampling of one thousand people who purportedly voted at
the polls in the last election, ask them if they actually voted, and confirm the
percentage who are valid voters. Researchers should conduct the survey
soon after an election to locate as many legitimate voters with fresh memories as possible.
Because many respondents would perceive voting as a social good,
some who did not vote might claim that they did, which may underestimate
the extent of fraud. A surveyor might mitigate this skew through the framing
of the question ("I've got a record that you voted. Is that true?"). Further,
some voters would not be located by researchers and others would refuse to
these
talk to researchers. Photo-identification proponents might construe
fraud." 7
nonrespondents as improper registrants that committed voter
Instead of surveying all voters to determine the amount of fraud, researchers might reduce the margin of error by focusing on a random
sampling of voters who signed affidavits in the states that request photo
identification but also allow voters to establish their identity through affidavit-Louisiana"8 and South Dakota." 9 In South Dakota, for example, only
2% of voters signed affidavits to establish their identity. If the survey indicates that 95% of those who signed affidavits are legitimate voters (and the
116.
Professor Lorraine C. Minnite argues as follows:
As a political scientist who has studied voter fraud I can tell you there are no reliable, officially
compiled national or even statewide statistics available on voter fraud .... Researchers working on voter fraud must construct their own datasets by culling information about allegations,
investigations, evidence, charges, trials, convictions, acquittals and pleas from local election
boards and local D.A.'s, county by county and sometimes town by town across the U.S. The
task is painstaking which explains in part why nobody has done it yet. Such a dataset is desirable because hard data are persuasive, at least with reasonable people.
Posting of Lorraine C. Minnite, [email protected], to [email protected] (Apr.
21, 2005) (on file with author).
117.
Id.
118.
LA. REv. STAT.
119.
S.D. CODIFIED LAWS
ANN.
§ 18:562 (Supp. 2006).
§ 12-18-6.2 (2004).
Michigan Law Review
[Vol. 105:631
other 5% were shown to be either fraudulent or were nonresponsive), this
suggests that voter fraud accounts for, at the maximum, 0.1% of ballots cast.
The affidavit study, however, is limited to two states, and it is unclear
whether this sample is representative of other states. 20 The affidavit study
also reveals information about the amount of fraud in a photo-identification
state with an affidavit exception-more voter fraud may exist in a state that
does not request photo identification. Further, the affidavit study fails to cap-
ture fraudulent voters without photo identification who left the polls without
voting when they were offered an affidavit to sign.
3. Examining Death Rolls
A comparison of death rolls to voting rolls might also provide an esti-
mate of fraud. Imagine that one million people live in state A, which has no
documentary-identification requirement. Death records show that twenty
thousand people passed away in state A in 2003. Cross-referencing this list
to the voter rolls shows that ten thousand of those who died were registered
voters, and that their names remained on the voter rolls during the Novem-
ber 2004 election. Researchers should look at what percentage of the ten
thousand dead-but-registered people "voted" in the November 2004 election. A researcher should distinguish the votes cast in the name of the dead
at the polls from those cast absentee (which could still be cast under a
photo-identification requirement). 2 ' This number would be extrapolated to
the electorate as a whole.
This methodology also has its strengths and weaknesses. If fraudulent
voters target the dead, the study might overestimate the fraud that exists
among living voters (although a low incidence of fraud among deceased
voters might suggest that fraud among all voters is low). The appearance of
fraud might also be inflated by false positives produced by a computer
match of different people with the same name. t22 Photo-identification advocates would likely assert that the rate of voter fraud could be higher from
fictitious names registered and that the death record survey would not cap120. The difficulty may be magnified in Louisiana in the aftermath of Hurricane Katrina's
displacement of hundreds of thousands of voters.
121.
Cf Jingle Davis, Even Death Can't Stop Some Voters, ATLANTA J.-CONST., Nov. 6, 2000,
at IA (finding that of 1.1 million deaths since 1980, 5412 ballots were cast in the name of dead
people over a twenty-year period, although not computing the fraud rate in relation to the total number of dead people who remained on the rolls between 1980 and 2000; asserting only that "the
actual number of ballots cast by the dead is fairly small"; and not distinguishing absentee votes from
those cast at the polls). In September 2005, in a contested Tennessee state senate seat in which Democrat Ophelia Ford won by 13 votes out of the 8653 votes cast in favor of the Democratic and
Republican candidates combined, an investigation showed that two votes had been cast by dead
people. See Lawrence Buser & Richard Locker, Senate Gets Nodfor Ford Vote Today, COM. APPEAL
(Memphis, Tenn.), Apr. 19, 2006, at Al; Marc Perrusquia, Dead Voter Evidence Goes to DA, COM.
APPEAL (Memphis, Tenn.), May 19, 2006, at B1.
122.
BRENNAN CTR. FOR JUST. AT N.YU. SCH. OF L. & MICHAEL P. McDONALD, ANALYSIS
OF THE SEPTEMBER 15, 2005 VOTER FRAUD REPORT SUBMITTED TO THE NEW JERSEY ATTORNEY
GENERAL 4 (2005). Any computer "matches" would require more detailed investigation to ensure
that they are not false positives. See id.
February 2007]
Voter Identification
ture that type of fraud because fictitious names registered would not show
up in the death records. Nevertheless, this study, combined with the other
two, would provide important insight into the magnitude of fraud likely to
exist in the absence of a photo-identification requirement.
B. Toward Better Data on Legitimate Voters
Excluded by Photo Identification
In addition to better data on fraud, policy-makers need better data regarding the impact of photo-identification requirements on participation by
legitimate voters before adopting the proposal. Scholars have defined citizen
participation as "purposeful activities in which citizens take part in relation
to government."'' Participation is a crucial democratic value. As Justice
Brandeis remarked, "the greatest menace to freedom is an inert people."' 24
Widespread participation serves four functions. First, it exposes decision-makers to a variety of ideas and viewpoints, which ensures fully
informed decisions. The failure to consider a wide, representative range of
views impairs deliberation. Second, widespread participation allows the
people, as a whole, to check the power of government officials who might
otherwise enact or tolerate abusive practices. 2 6 Accountability to the electorate as a whole ensures democratic legitimacy, 12 which in turn may
increase the likelihood that citizens will voluntarily comply with such decisions.128 Third, widespread participation allows for a redistribution of
government resources and priorities to reflect evolving problems and
needs. 129 Finally, widespread participation furthers self-fulfillment and selfdefinition of individual
citizens who play a role in shaping the decisions that
130
affect their lives.
123.
Stuart Langton, What is Citizen Participation?,in CITIZEN
PARTICIPATION IN AMERICA
13, 17 (Stuart Langton ed., 1978) (emphasis omitted).
124.
Whitney v. California, 274 U.S. 357, 375 (1927) (Brandeis, J., concurring).
125. See Nancy Perkins Spyke, Public Participationin Environmental Decisionmaking at the
New Millennium: Structuring New Spheres of Public Influence, 26 B.C. ENVTL. AFF. L. REV.263,
267-68 (1999).
126. Cf THE FEDERALIST No. 51, at 349 (James Madison) (Jacob E. Cooke ed., 1961) ("A
dependence on the people is no doubt the primary controul [sic] on the government .... ).
127. See DON HERZOG, HAPPY SLAVES: A CRITIQUE OF CONSENT THEORY 205-07 (1989):
Hanna Fenichel Pitkin, The Concept of Representation, in REPRESENTATION 1, 22 (Hanna Fenichel
Pitkin ed., 1969); Bernard Manin, On Legitimacy and Political Deliberation, 15 POL. THEORY 338,
357 (Elly Stein & Jane Mansbridge trans., 1987).
128. See MARY GRISEz KWEIT & ROBERT W. KWEIT, IMPLEMENTING CITIZEN PARTICIPATION IN A BUREAUCRATIC SOCIETY: A CONTINGENCY APPROACH 132 (1981) ("The more satisfied
the citizens are with participation, the more trusting and efficacious they will be."); Luis FuentesRohwer, The Emptiness of Majority Rule, I MICH. J. RACE & L. 195, 201 (1996) ("To deserve the
democratic denomination, the people must take part in political affairs.").
129. See KWEIT & KWEIT, supra note 128, at 33-34 (asserting that the goals of public participation include the redistribution of power).
130. See Frank 1. Michelman, Conceptions of Democracy in American ConstitutionalArgument: Voting Rights, 41 FLA. L. REV. 443, 451 (1989) (discussing a "constitutive" vision of politics
whereby citizens define themselves through their participation); see also C.B. MACPHERSON, THE
Michigan Law Review
[Vol. 105:631
Even in the absence of a photo-identification requirement, the United
States already has one of the lowest voter-participation rates among the
world's democracies. We trail many other established and developing democracies in voter turnout by twenty to thirty percentage points, and one
survey ranked the United States 138th of 170 democracies.' In light of the
importance of widespread participation, policy-makers should examine the
data on the number of legitimate voters a photo-identification requirement
would exclude.
A driver's license is the most common form of state-issued photo identification. The 2005 Carter-Baker Commission estimated that 12% of votingage Americans lack a driver's license, and an analysis of 2003 Census and
Federal Highway Administration data estimates that twenty-two million
voting-age citizens lack a driver's license. 33 Some 3-4% of voting-age
Americans carry a nondriver's photo-identification card issued by a state
motor vehicle agency in lieu of a driver's license.'T 34 Thus, according to the
2001 Carter-Ford Commission, an estimated 6-10% of voting-age Americans (approximately eleven million to twenty million potential voters) do
license or a state-issued nondriver's photonot possess a driver's
135
identification card.
Federal data suggests that younger and older Americans are less likely
to have a driver's license. While the rate of unlicensed individuals ages
twenty-five to sixty-nine hovered between 5% and 11% in 2003, the percentages of older and younger Americans who lack a driver's license were
much higher.
LIFE AND TIMES OF LIBERAL DEMOCRACY 47-48, 51-52 (1977) (asserting that public participation
increases "the amount of personal self-development of all the members of society").
131.
RAFAEL L6PEZ PINTOR ET AL., INT'L INST. FOR DEMOCRACY & ELECTORAL ASSISTANCE,
VOTER TURNOUT SINCE 1945: A GLOBAL REPORT 83--84 (2002), http://www.idea.int/publications/
vt/upload/VTscreenopt_2002.pdf.
132. See COMM'N ON FED. ELECTION REFORM, supra note 4, at 73 n.22. According to the
Federal Highway Administration, 13.2% of U.S. residents sixteen years and older lacked a driver's
license. See FED. HIGHWAY ADMIN., U.S. DEP'T OF TRANSP., LICENSED DRIVERS BY SEX AND RATIO
TO POPULATION -2003 (2004), http://www.fhwa.dot.gov/policy/ohim/hs03/pdf/dllc.pdf.
133.
BRENNAN CTR. FOR JUST. & OVERTON, supra note 12, at 24 n.9.
Publius, supra note 11, at 289 (citing FED. ELECTIONS COMM'N, THE IMPACT OF THE
NATIONAL VOTER REGISTRATION ACT OF 1993 ON THE ADMINISTRATION OF ELECTIONS FOR FEDERAL OFFICE 5-6 (1995-96)).
134.
135.
See TASK FORCE ON THE FED. ELECTION SYS., supra note 12, ch. 6, at 4; see also BRENNAN CTR. FOR JUST., supra note 14 (reporting on survey indicating that eleven percent of U.S.
citizens lack current, unexpired government-issued photo identification); Weinschenk v. Missouri,
No. SC88039, 2006 Mo. LEXIS 122, at *7 (Mo. Nov. 7, 2006) (indicating that, based on separate
analyses of the Missouri Secretary of State and the Missouri Department of Revenue, between three
and four percent of Missouri citizens lacked the requisite photo identification to vote).
Voter Identification
February 20071
U.S.
RESIDENTS UNLICENSED BY AGE"'
%w/o
%w/o
Age
18
19
20
21
license
32.5
26
22.9
20.6
1
22
20.1
23
24
18.1
14.5
Age
70-74
75-79
80-84
85+
license
14.3
18.6
26.9
48.3
Other studies on demographic disparities in photo identification focus
largely on particular areas and localities. According to the Georgia chapter
of the American Association of Retired Persons, for example, 36% of Georgians over age seventy-five lack a driver's license.'37 In 1994, the U.S.
Department of Justice found that African-Americans in Louisiana were four
to five times less likely than white residents to have government-sanctioned
photo identification.' Of the forty million Americans with disabilities,
nearly 10% lack identification issued by the government.'3 9
One of the more comprehensive studies was completed in June 2005 by
the Employment and Training Institute at the University of WisconsinMilwaukee. The study used census data and data from the Wisconsin Department of Transportation computer database for licensed drivers, and it
found that senior citizens, younger people, and people of color were less
likely to possess a driver's license.' 40 The study determined that 23% of Wisconsin residents (177,399 individuals) over age sixty-five do not have a
Wisconsin driver's license or state photo identification. 14 ' 30% of voting-age
residents in Milwaukee County lack a driver's license, compared with 12%
of residents in the balance of Wisconsin. Statewide, significant racial and
136. See FED. HIGHWAY ADMIN., supra note 132; see also BRENNAN CTR. FOR JUST., supra
note 14 (reporting on survey indicating that eighteen percent of U.S, citizens aged 18-24 and eighteen percent of individuals aged 65 and above lack current, unexpired government-issued photo
identification).
137. Nancy Badertscher & Tom Baxter, State AARP Criticizes Voter ID Bill,
CONST., Mar. 17, 2005, at 4C.
ATLANTA J.-
138. See Letter from Deval L. Patrick, Assistant Att'y Gen., Civil Rights Div., U.S. Dep't of
Justice, to Sheri Marcus Morris, La. Assistant Att'y Gen. (Nov. 21, 1994).
139. Ctr. for Policy Alternatives, Voter Identification and Integrity, http://www.cfpa.org/issues/
issue.cfm/issue/VoterlDandlntegrity.xml (last visited Sept. 10, 2006).
140.
PAWASARAT,
141.
Id. at 1.
supra note 14, at 1-2.
142. Id. at 6. According to Census estimates, the voting-age population of Milwaukee County
consists of 425,372 residents who reside in the city of Milwaukee and 268,667 who live in suburban
communities. Id. at 15. In New York City, up to three million registered voters lack a driver's license. BRENNAN CTR. FOR JUST. & OVERTON, supra note 12, at 24 n.9 (citing Elizabeth Daniel, The
New Voter Identification Requirement, GOTHAM GAZETTE (N.Y.), Apr. 2002, http://
www.gothamgazette.com/article//20020401/17/728).
Michigan Law Review
[Vol. 105:631
age disparities also exist, the most striking being that 78% of African43
American males ages eighteen to twenty-four lack a valid driver's license.
RACE AND PERCENTAGE OF UNLICENSED WISCONSIN RESIDENTS' 44
White Males
White Females
Black Males
Black Females
Latino Males
Latino Females
Ages 18-24
36
25
78
66
57
63
All Voting Ages
17
17
55
49
46
59
The data above suggests that a photo-identification requirement would
exclude some legitimate voters and would have a disparate demographic
impact. '41
A photo-identification requirement may not exclude as many voters,
however, as the numbers initially suggest. Assuming that those without
photo identification are disproportionately poor and have lower voter participation rates, the percentage of those who lack photo identification may
be lower among the electorate than it is among the entire voting-age popula143.
PAWASARAT, supra note 14, at 5.
144. Id. at 4, 5; see also BRENNAN CTR. FOR JUST., supra note 14 (reporting on survey of
voting-age U.S. citizens indicating that fifteen percent of those who earn less than $35,000 per year
and twenty-five percent of African-Americans lack current, unexpired government-issued photo
identification).
145. Political appointees in the U.S. Justice Department recently used skewed data to suggest
that photo-identification requirements have no adverse impact on voters of color. In a letter to U.S.
Senator Christopher Bond explaining the Justice Department's rationale in failing to object to Georgia's new photo-identification law, Assistant Attorney General William E. Moschella asserted that
previous identification requirements had not diminished African-American turnout in the 2000 or
2004 general elections. Letter from William E. Moschella, Assistant Att'y Gen., U.S. Dep't of Justice, to Christopher S. Bond, U.S. Sen. (Oct. 7, 2005), http://www.usdoj.gov/crt/voting/nisc/
gajid bondjtr.pdf. Political factors unrelated to voter identification rules, however-such as mobilization efforts by parties, controversial issues, and a polarized electorate-may increase turnout in a
later contest. Further, the earlier identification laws were not photo-identification requirements, but
much less restrictive practices that allowed voters to establish their identity by using any one of
seventeen types of documentary identification (including nonphoto identification such as utility bills
or bank statements) or by signing an affidavit. Help America Vote Act requirements, which applied
only to first-time voters who registered by mail, allowed voters to establish their identity through
nonphoto documentary identification such as utility bills or bank statements.
Assistant Attorney General Moschella also claimed to rely on Georgia motor vehicle administration data that suggested that African-Americans were slightly more likely to possess
identification than whites. Id. This data is inconclusive, however, because Georgia provided racial
data for less than sixty of those with identification, and there is no evidence that this pool is a statistically representative sampling of voters from across the state. Indeed, county data suggests the
opposite. The ten Georgia counties with the highest percentage of African-Americans (59.5-to
77.8% black) have only 87.7% of the identification cards per one thousand voting-age residents as
the ten counties with the highest percentage of whites (93.4 to 97.1% white). See Letter from Adam
Cox, Professor, Univ. of Chicago Law Sch.; Heather Gerken, Professor, Harvard Law Sch.; Michael
Kang, Professor, Emory Law Sch.; Spencer Overton, Professor, The George Washington Univ. Law
Sch.; and Daniel Tokaji, Professor, Ohio State Univ. Moritz Coll. of Law, to John Tanner, Assistant
Att'y Gen., Civil Rights Div., U.S. Dep't of Just. (Aug. 19, 2005) (on file with author).
February 20071
Voter Identification
tion. Further, the most restrictive existing laws (in Georgia and Indiana) allow voters to establish their identity using a U.S. passport or federal and
state employee photo-identification card, and some voters who lack a
6
driver's license will possess one of these documents.' Also, in the aftermath of the terrorist attacks of 9/11, the number of individuals who do not
have photo identification may drop as Americans find that it is even more
47
difficult to function in modem life without a photo-identification card.'
Finally, if a photo-identification requirement to vote is enacted, some people
who lack state-issued photo identification will likely obtain it so that they
can vote (although the percentage who will do so remains unclear).
Other factors suggest that a photo-identification requirement could exclude many more than the six to ten percent of the voting-age population
who lack state-issued photo identification 4 and that demographic disparities may be greater. Some legitimate voters who have been issued a driver's
license or other identification may not bring it to the polls because the card
was stolen, lost, or simply forgotten. Further, the numbers of individuals
without valid photo identification may rise due to the heightened burdens of
the REAL ID Act. After 2007, the REAL ID Act prohibits states from issuing a driver's license or nondriver's identification card unless a person
presents documentary proof of (a) her full legal name and date of birth, (b)
her Social Security number (or the fact that she is not eligible49 for one), (c)
citizenship.
the address of her principal residence, and (d) her
A law that requires a voter's current address to appear on the photo50
identification card would also drive up the number of those excluded. The
University of Wisconsin-Milwaukee study confirmed that transient populations were less likely to have valid driver's licenses. Of the 12,624 students
living in residence dorms at Marquette University, the University of Wisconsin-Madison, and the University of Wisconsin-Milwaukee, less than 3%
had driver's licenses that listed their dorm's address.'' Over 76% of Wisconsin renters moved between January 1995 and March 2000, compared
with only 22% of homeowners. 1 2 During this same time period, 44% of
whites moved, compared with 75% of Asian-Americans, 74% of Latinos,
3'
63% of African-Americans, and 61% of Native-Americans.'
146. The Carter-Baker Commission's recommendation limited acceptable forms of identification to be issued under the Real ID Act to a driver's license or state-issued photo identification.
COMM'N ON FED. ELECTION REFORM, supra note 4, at 19.
147.
See id. at 21.
148.
See infra notes 132-135.
REAL ID Act of 2005,49 U.S.C.A. § 30301 (Supp. 2006).
150. For example, proposed legislation in Ohio indicated that a photo-identification card must
include a voter's current address (this provision was later removed). Daniel P. Tokaji, Ohio Election
Bill Clears Senate, ELECTION LAW @ MORITZ, Dec. 15, 2005, http://moritzlaw.osu.edu/
blogs/tokaji/2005/12/ohio-election-bill-clears-senate.html (last visited Sept. 10, 2006).
PAWASARAT, supra note 14, at 11-12.
151.
149.
152.
Id. at 17.
153.
Id. at 18.
Michigan Law Review
[Vol. 105:631
Rather than rely on uninformed "hunches," such as the assumption that
the terrorist attacks of 9/11 will significantly increase the number of Americans who possess identification, more detailed empirical work is needed to
determine the extent to which a photo-identification requirement will shape
the electorate. What percentage of the electorate (rather than the general
population), for example, lacks a state-issued photo-identification card?
What percentage of those who have been issued photo identification will fail
to bring it to the polls?
Some answers may come from data on affidavits in states that allow voters without photo identification to affirm their identity under penalty of
perjury. Affidavits provide insight into the percentage of Americans who fail
to bring either a license or some other form of photo identification to the
polls.
As mentioned earlier, South Dakota and Louisiana request photo identification but allow voters to sign an affidavit in lieu of presenting such
identification. The number and demographic patterns of the affidavits in
these states could indicate which voters would be excluded if a photoidentification card were an absolute requirement to vote. For example, reports of the 2004 primary in South Dakota showed that 2% of voters used an
affidavit statewide, whereas between 4% and 16% of voters used affidavits
in the predominantly 54Native-American counties of Shannon, Todd, Corson,
Dewey, and Zieback.
Affidavit data is important, but not determinative. Affidavit data may
underestimate the number of people who lack photo identification. For example, the affidavit records would not record the legitimate voter who lacks
photo identification and does not cast a ballot because (1) the pollworker did
not offer an affidavit to the voter or (2) the affidavit process was much more
time-consuming and the voter decided not to wait. On the other hand, the
affidavit does not measure voters who would obtain a photo-identification
card if it were an absolute requirement for voting, and a collection of affidavits may include forms completed by some fraudulent individuals who
154. Chet Brokaw, Lawmakers asked to repeal voter identification law, ASSOCIATED PRESS,
Jul. 15, 2004. Political appointees at the Justice Department have recently refused to examine affidavit evidence in reviewing whether Georgia's photo-identification law disproportionately excluded
people of color. In a letter that I drafted with other law professors before the Justice Department
precleared the Georgia identification requirement, we asked officials to request and review affidavit
information before making a decision. Specifically, we wrote as follows:
Indeed, the ultimate question is not whether state records show that minorities are just as likely
as whites to have applied for a driver's license or other govemment-issued ID. The most important question is what minorities bring to the polls on Election Day to establish their identity.
On that score, Georgia has failed to satisfy its burden by providing the most relevant information-racial data on those who have utilized the affidavit ID option.
Letter from Adam Cox et al. to John Tanner, supra note 145. Assistant Attorney General William E. Moschella characterized the request for affidavit information as suggesting that "the
Department seek data to establish that racial minorities may be more likely than non-minorities to
misplace or forget their identification when coming to the polls. Such a notion is incredibly demeaning to minorities, and this Department emphatically declines to entertain such a request." Letter
from William E. Moschella to Christopher S. Bond, supra note 145.
February 2007]
Voter Identification
forged the signatures of others (although the study of fraud proposed in Section III.A may address this issue).
While partisans can construe any study to favor their preferred outcomes, policy-makers should obtain and consider the best data available.
Granted, empirical data is sometimes misleading due to value-driven research assumptions or deliberate skewing or manipulation of data.'55 Even
for those who act in good faith, it may also be difficult to separate empirical
data from normative democratic values in assessing and managing the risks
of voter fraud and the exclusion of legitimate voters by a photoidentification requirement.
Rather than using these shortcomings as a justification to dismiss empirical data completely and defer solely to misleading anecdotes and flawed
analogies, policy-makers should acknowledge the limitations of empirical
study, scrutinize research methodologies, and make informed decisions
based on more information rather than less. Empirical data is not perfect,
but it allows for a better understanding of the true costs and benefits of a
photo-identification requirement and permits a more honest debate about the
democratic values at issue.
IV. THE LEGAL
STATUS OF PHOTO-IDENTIFICATION REQUIREMENTS
Empirical data is crucial not just for policy-making but also for analyzing whether proposed photo-identification requirements comply with
constitutional and statutory requirements. Empirical evidence allows courts
to determine whether photo-identification requirements constitute an undue
burden on the fundamental right to vote, a poll tax, or a violation of Section
2 of the Voting Rights Act."'
The importance of ascertaining the empirical magnitude of voter fraud
and the suppression of legitimate votes was recently emphasized in a challenge to Arizona's requirement that voters produce photo identification or
two pieces of nonphoto documentary identification. While the U.S. Supreme
Court expressed "no opinion ... on the correct disposition ... or on the ultimate resolution" of the case and vacated a lower court injunction on
procedural grounds, it emphasized that "the facts in these cases are hotly
contested" and noted the need for judges to take adequate "time to resolve
155.
See, e.g., Wendy E. Wagner, Congress, Science, and Environmental Policy, 1999 U. ILL.
L. REV. 181 (arguing against overreliance on "science" in agency decision-making).
156. Empirical data can help courts properly evaluate whether photo-identification requirements violate other legal provisions, such as the Twenty-sixth Amendment (voting rights of citizens
eighteen years of age or older shall not be denied or abridged on account of age), U.S. CoNsT.
amend. XXVI, the Fifteenth Amendment (prohibiting racial discrimination in voting), id. at amend.
XV, the Voting Accessibility for the Elderly and Handicapped Act, 42 U.S.C. §§ 1973ee to 1973ff-6
(2000), and various state constitutional claims.
Michigan Law Review
[Vol. 105:631
the factual disputes."'57 In a separate concurrence, Justice Stevens was more
direct about the empirical data needed to decide the matter:
At least two important factual issues remain largely unresolved: the scope
of the disenfranchisement that the novel identification requirements will
produce, and the prevalence and characterof the fraudulent practices that
allegedly justify those requirements.'
A. Burdening the FundamentalRight to Vote
Depending on the amount of voter fraud that exists and the number of
legitimate voters who would be excluded, a photo-identification requirement
may unduly burden the fundamental right to vote that stems from the First
and Fourteenth Amendments.'59
While allowing that "there must be a substantial regulation of elections
if they are to be fair and honest,''t6 courts use the following test to determine
whether an election procedure unduly abridges the right to vote:
[A] court must resolve such a challenge by an analytical process that parallels its work in ordinary litigation. It must first consider the character and
magnitude of the asserted injury to the rights protected by the First and
Fourteenth Amendments that the plaintiff seeks to vindicate. It then must
identify and evaluate the precise interests put forward by the State as justifications for the burden imposed by its rule. In passing judgment, the Court
must not only determine the legitimacy and strength of each of those interests, it also must consider the extent to which those interests make it
necessary to burden the plaintiff's rights. Only after weighing all these factors is the reviewing court in a position to decide whether the challenged
provision is unconstitutional .
157.
Purcell v. Gonzalez, 127 S. Ct. 5, 8 (2006) (per curiam).
158.
Id. (Stevens, J., concurring) (emphases added).
159. See Bush v. Gore, 531 U.S. 98, 104 (2000) (per curiam) ("When the state legislature
vests the right to vote for President in its people, the right to vote as the legislature has prescribed is
fundamental... ");Ill. State Bd. of Elections v. Socialist Workers Party, 440 U.S. 173, 184 (1979)
("[V]oting is of the most fundamental significance under our constitutional structure."); Reynolds v.
Sims, 377 U.S. 533, 561-62 (1964) ("Undoubtedly, the right of suffrage is a fundamental matter in a
free and democratic society."); Wesberry v. Sanders, 376 U.S. 1, 17-18 (1964) ("No right is more
precious in a free country than that of having a voice in the election of those who make the laws
under which, as good citizens, we must live. Other rights, even the most basic, are illusory if the
right to vote is undermined. Our Constitution leaves no room for classification of people in a way
that unnecessarily abridges this fight."). Advocates of an amendment to the U.S. Constitution that
calls for an explicit fight to vote note that some U.S. Supreme Court Justices have observed that no
such right exists. See, e.g., Jesse Jackson, Editorial, No Change in No-Account System, CHt. SUNTM Es, Nov. 23, 2004, at 37 (citing Bush v. Gore, 531 U.S. 98, 104 (2000)). This Article does not
address the need for the passage of a right-to-vote constitutional amendment, but it does note that
the U.S. Supreme Court has repeatedly recognized that voting is a fundamental right that arises from
the First and Fourteenth Amendments.
160.
Storer v. Brown, 415 U.S. 724, 730 (1974).
161. Anderson v. Celebrezze, 460 U.S. 780, 789 (1983); accord Timmons v. Twin Cities Area
New Party, 520 U.S. 351, 358 (1997); Burdick v. Takushi, 504 U.S. 428, 434 (1992).
February 2007]
Voter Identification
The test operates on a continuum-there exists "[n]o bright line" that
separates permissible from unconstitutional election regulation.
62
63
Strict
scrutiny applies to "severe" restrictions on voting rights, lesser burdens
trigger less exacting review, and "reasonable, nondiscriminatory restric'
tions" are usually constitutional if "important regulatory interests" exist. 6
1. Assessing the Voters' Burden Relative to the State Interest
Hard data is especially valuable in assessing the burdens of a photo-
identification requirement on voters and the state's interest in preventing
fraud. Without hard data, judges would likely engage in ad hoc, contestable
conjecture about the danger of fraud and the difficulty of obtaining a photoidentification card. Many judges inclined to favor a photo-identification re-
quirement, for example, can invoke a plausible anecdote of fraud or use
flowery language to proclaim that photo identification is necessary to maintain voter confidence. These judges can speculate that photo identification
162.
immons, 520 U.S. at 359 ("No bright line separates permissible election-related regulation from unconstitutional infringements on First Amendment freedoms."); Anderson, 460 U.S. at
789-90 ("The results of this evaluation will not be automatic; as we have recognized, there is 'no
substitute for the hard judgments that must be made.' " (quoting Storer, 415 U.S. at 730)); Storer,
415 U.S. at 730 ("[N]o litmus-paper test [I separatfes] those restrictions that are valid from those
that are invidious ....
"(alteration in original)).
163.
Burdick, 504 U.S. at 434; Norman v. Reed, 502 U.S. 279, 289 (1992).
164.
Timmons, 520 U.S. at 358-59.
165. See, e.g., League of Women Voters v. Blackwell, 340 F. Supp. 2d 823, 829 (N.D. Ohio
2004) ("If elections are not substantially free from fraud and other irregularities, public confidence
in their integrity and the validity of their results, which is essential to the maintenance of ordered
liberty, is threatened."). Some photo-identification advocates argue that regardless of the magnitude
of fraud, "the perception of possible fraud contributes to low confidence in the system' and that a
photo-identification requirement "can enhance confidence' COMM'N ON FED. ELECTION REFORM,
supra note 4, at 18-19. Growing cynicism in the absence of a photo-identification requirement, the
argument goes, lowers voter participation. See FUND, supra note 22, at 2 (suggesting that low confidence may result in low voter turnout); Purcell v. Gonzalez, 127 S.Ct. 5, 7 (2006) (per curiam)
("Voter fraud drives honest citizens out of the democratic process and breeds distrust of our government. Voters who fear their legitimate votes will be outweighed by fraudulent ones will feel
disenfranchised.").
Several leading scholars have criticized the Court's dicta in Purcell regarding the perception of
fraud. See Alex Keyssar, "Disenfranchised"? When Words Lose Meaning, THE HUFFINGTON POST,
Oct. 22, 2006, http://www.huffingtonpost.com/alex-keyssar/disenfranchised-when-_b_32241.html
("FEEL disenfranchised? Is that the same as 'being disenfranchised'? So if I might 'feel' disenfranchised, I have a right to make it harder for you to vote? What on earth is going on here?"); Richard
L. Hasen, Election Deform: The Supreme Court Messes Up Election Law, SLATE, Oct. 24, 2006,
http://www.slate.com/id/2152116/ ("[T]he Supreme Court's statement in Purcell ...confuses the
empirical analysis at the heart of these cases by suggesting, without any proof whatsoever, that
concerns about voter fraud '[drive] honest citizens out of the democratic process and [breed] distrust
of our government."'); Michael C. Dorf, In a Brief Unsigned New Opinion, The Supreme Court
Sends the Wrong Signal on Voter ID and Voter Fraud, FINDLAw, Nov. 6, 2006,
http://writ.news.findlaw.com/dorf/20061106.html (noting that "the Supreme Court asserts that fears
of voter fraud will lead other voters to feel disenfranchised ... without citing any supporting evidence from the record" and that Purcell suggests "that this interest in preventing feelings of
disenfranchisement rises to a level sufficient to override the voting rights claims of people who will
actually be disenfranchised.").
A lack of empirical data also allows photo-identification proponents to make claims about the
perception of fraud without explaining the extent to which a lack of a photo-identification
Michigan Law Review
[Vol. 105:631
is not unreasonably burdensome because of fee waivers and new photoidentification distribution programs.' 66 A judge skeptical of a photoidentification requirement, on the other hand, can underemphasize the existence of voter fraud and overemphasize anecdotes about individuals who
had difficulties securing a photo-identification card.
Reliance on these personal assumptions allows for the charge that personal political ideology rather than law shaped the judge's holding. In light
of the political nature of the photo-identification debate, the institutional
limitations of courts, and the important democratic values furthered by both
widespread participation and the prevention of voter fraud, judges should
look to empirical data for more reasoned analysis and consistency in decision making.
Imagine, for example, a state in which about one million citizens regularly turn out to vote. Empirical studies suggest that five percent of
legitimate voters in the state (fifty thousand people) would not bring a
photo-identification card to the polls if it were required, and most of these
voters would be ethnic minorities who regularly support Party A. Studies
also suggest that in the absence of a photo-identification requirement at the
polls, fifty fraudulent votes would be cast (0.005% of votes cast).
In considering the magnitude of the injury, 167 the court can look to the
evidence that suggests fifty thousand legitimate voters will not cast a ballot
because of the photo-identification requirement. The disproportionate im-
pact of the proposal on ethnic minorities who vote for Party A suggests that
the restriction lacks neutrality.16 In examining the "legitimacy and
strength"'' 69 of the "precise interests put forward by the State,"'"7 the court
can quantify the state's interest in preventing fifty fraudulent votes. The
court can determine whether it is "necessary to burden"'' 7' the legitimate vot-
ers with a photo-identification requirement by looking at data on the
requirement-relative to the existence of other factors, such as manipulation of voting rules by
politicians that suppresses voter turnout-lowers voter confidence and participation. Cf Nathaniel
Persily & Kelli Lammie, Perceptions of Corruption and Campaign Finance: When Public Opinion
Determines ConstitutionalLaw, 153 U. PA. L. REV. 119 (2004) (asserting that popular perceptions
of corruption are related to factors other than campaign finance laws and that restrictive campaign
reforms would not lower the perception of corruption, and concluding that courts should not base
their decisions about the need for campaign restrictions on popular opinion);
Robert Bauer, Judicial "Wisdom," MORE SoFr MONEY HARD LAW, Oct. 23, 2006, http://
www.moresoftmoneyhardlaw.com/updates/electionadministration.html?AID=845
(observing that
by invoking the "fear" of fraud, advocates "bypass the demands of evidence").
166. Ind. Democratic Party v. Rokita, No. I:05-CV-0634-SEB-VSS, 2006 WL 1005037, at
*36 (S.D. Ind. Apr. 14, 2006) ("[T]he individuals and groups that Plaintiffs contend will be disproportionately impacted by [the statute] all appear fully capable of availing themselves of the law's
exceptions so that they do not need to obtain photo identification in order to vote.").
167. Anderson, 460 U.S. at 789.
168.
See id.
169.
Id.
170.
Id.
171.
Id.
February 2007]
Voter Identification
effectiveness of alternatives, such as an affidavit, in deterring most fraudu-
lent voters and very few legitimate ones."
2. Tailoring
A court should also use empirical information to determine whether a
photo-identification requirement is properly tailored. To be properly tailored, a statute must further the government's objectives, must not be
overinclusive or underinclusive to an unacceptable extent, and must not be
unnecessarily burdensome." 3 A statute is overinclusive when the proportion
of invalid applications of the statute is substantially high relative to the proportion of valid applications.7 4 A statute is underinclusive when it fails to
prevent a relatively large number of activities that pose the danger that the
statute was designed to prevent.
The tailoring requirement addresses the difficulty in crafting a single,
bright-line voter-identification law that would prevent all voter fraud while
simultaneously including all legitimate voters. Any rule will tend to be ei-
ther underinclusive and allow for some fraudulent voting, or overinclusive
and inhibit some legitimate votes, and often both.
The amount of overinclusiveness and underinclusiveness that a court
should tolerate depends on the level of scrutiny. As mentioned above, the
appropriate level of scrutiny to apply to a photo-identification requirement
depends on the magnitude of the burden relative to the precise interest of the
state. Regulations subject to strict scrutiny must be narrowly tailored to advance a compelling state interest, and as much as possible they should avoid
restricting constitutionally protected activity that does not pose the danger
that motivated the regulation. 175 Regulations subject to intermediate scrutiny
must be substantially related to an important government interest, 1 6 and
172. See infra Part V for a discussion of supplements and alternatives to absolute photoidentification requirements.
173. See Eugene Volokh, Freedom of Speech, Permissible Tailoring and Transcending Strict
Scrutiny, 144 U. PA. L. REV. 2417, 2422-23 (1996).
174. See New York v. Ferber, 458 U.S. 747, 773 (1982) (rejecting a substantial overbreadth
claim because the statute's "legitimate reach dwarfs its arguably impermissible applications");
Broadrick v. Oklahoma, 413 U.S. 601, 615 (1973) ("[W]e believe that the overbreadth of a statute
must not only be real, but substantial as well, judged in relation to the statute's plainly legitimate
sweep."). Scholars have described applications of the narrow tailoring and First Amendment substantial overbreadth tests as substantively identical. See Richard L. Hasen, Measuring Overbreadth:
Using EmpiricalEvidence to Determine the Constitutionalityof Campaign FinanceLaws Targeting
Sham Issue Advocacy, 85 MINN. L. REV. 1773, 1782 n.46 (2001); Henry Paul Monaghan, Overbreadth, 1981 SuP. CT. REV. 1,37 n.152.
175. See FEC v. Mass. Citizens for Life, Inc., 479 U.S. 238, 265 (1986)
ble, government must curtail speech only to the degree necessary to meet the
hand, and must avoid infringing on speech that does not pose the danger that
tion."); see also Denver Area Educ. Telcomms. Consortium, Inc. v. FCC, 518
(Kennedy, J., concurring in part, concurring in the judgment in part, dissenting
176.
E.g., United States v. Virginia, 518 U.S. 515, 524 (1996).
("Where at all possiparticular problem at
has prompted regulaU.S. 727, 807 (1996)
in part).
Michigan Law Review
[Vol. 105:631
regulations subject to rational basis scrutiny must be rationally related to a
177
legitimate government interest.
Whatever scrutiny is applied, data allows a judge to consider the number
of applications of the statute that pose the danger that the statute was designed to prevent (i.e., fraudulent votes) relative to the number of
applications covered by the statute that do not pose the danger the statute
was designed to prevent (i.e., legitimate voters who lack photo identification).
For example, assume data reveals that a photo-identification requirement
excluded one thousand votes, and that 990 of these were fraudulent and ten
were legitimate. This data provides strong evidence that such a photoidentification requirement is narrowly tailored.
In contrast, assume that 990 of the excluded votes were legitimate and
only ten were fraudulent. Further, assume that the regulation follows the
Carter-Baker Commission's proposal and requires photo identification at the
polls but merely a signature from absentee voters, and thus the regulation
tolerates three thousand fraudulent absentee ballots. This data suggests that
the regulation is at once so overinclusive and so underinclusive that it is not
rationally related to the state's purported interest in preventing fraud.
Granted, the magnitude of the burden, the appropriate level of scrutiny,
and the proper tailoring cannot be reduced to a simple mathematical formula. Different methodologies and underlying assumptions, along with
other variables, can result in varying numbers. Even when judges agree on
data, they will still harbor normative differences that might lead them to
vastly different conclusions. For example, judges might differ on whether a
photo-identification requirement that deters one thousand fraudulent votes
and one thousand legitimate votes is "narrowly tailored," or whether a
photo-identification law that deters 250 fraudulent votes and one thousand
legitimate votes is "rationally related" to the prevention of fraud. To some
judges, fraudulent votes taint democracy much more than reduced participation by legitimate voters. Other judges might err on the side of preserving
access and risk ten fraudulent votes to ensure that legislatures do not exclude a single legitimate voter (much as the "reasonable doubt" standard in
the criminal context in theory errs against convicting criminal defendants).
Another question involves whether photo identification requirements are
"reasonable" and "non-discriminatory." To the extent that the regulations
disproportionately exclude people of color, poorer Americans, disabled
Americans, young Americans, or senior citizens, how significant must this
demographic skew be before it becomes intolerable? How should judges
tackle the thorny problem of disproportionate exclusion if the data shows
that the rate of fraud is also disproportionately high among these voters?
What if one hundred fraudulent votes were cast and split evenly between the
political parties (both Republicans and Democrats received fifty fraudulent
177.
E.g., Lyng v. Int'l Union, 485 U.S. 360, 370 (1988).
February 2007]
Voter Identification
votes apiece), but a photo-identification requirement deterred voting by
ninety legitimate Democratic voters and no legitimate Republican voters?8
Statistical data will not answer these normative questions, but such data
is necessary to make an honest conversation about normative values possible. Absent data, judges and advocates can avoid a discussion of different
normative values by using assumptions, anecdotes, and analogies to paint a
factual picture that appears to support their desired outcome.
An "undue burden" legal analysis also requires that a court examine
whether less restrictive alternatives of voter identification exist. 79 This Article explores alternatives below in Part V.
B. Photo-IdentificationFees as Poll Taxes
Many states charge a fee to issue a photo-identification card,180 and better data can establish whether requiring a photo identification to vote
violates the Twenty-Fourth Amendment's prohibition on poll taxes. Georgia's law allowed residents to file an affidavit of indigency to receive free
state-issued photo identification. 8 2 In Common Cause/Georgia v. Billups,
however, the court found that "very few voters likely will take advantage of
the fee waiver affidavit option" due to embarrassment about their poverty, or
to being non-indigent and unwilling either to lie about financial status or to
178. Christopher S. Elmendorf observes that voter-identification laws may improperly "fence
out" voters because of the way they vote. Cf Carrington v. Rash, 380 U.S. 89, 94 (1965) ("'Fencing
out' from the franchise a sector of the population because of the way they may vote is constitutionally impermissible."). Elmendorf proposes that judges ascertain whether voter-identification laws
unconstitutionally "fence out" voters by evaluating whether:
(1) the voting restriction was enacted substantially along partisan lines; (2) there is some evidence that the law will disproportionately inconvenience citizens who are statistically more
likely to support the opposition party; and (3) the law is a permanent measure, rather than a
time-limited experiment with provisions for independent evaluation of its impacts on electoral
participation by the ostensibly disadvantaged classes.
Christopher S. Elmendorf, Burdick or Carrington: "Fencing Out" and the Voter ID Litigation,
ELECTIONLAW @MORITZ, Sept. 12, 2006, http://moritzlaw.osu.edu/electionlaw/comments/2006/
060912.php.
179.
Dunn v. Blumstein, 405 U.S. 330, 342-43 (1972).
180. See, e.g., Common Cause/Ga. v. Billups, 406 F. Supp. 2d 1326, 1369 (N.D. Ga. 2005)
("The fee for a Photo ID card [in Georgia] is $20 for a five-year card and $35 for a ten-year card.").
181.
U.S. CONST. amend. XXIV, § I ("The right of citizens of the United States to vote ...
shall not be denied or abridged by the United States or any State by reason of failure to pay any poll
tax or other tax."). The U.S. Supreme Court prohibited poll taxes in state and local elections when it
held that "a State violates the Equal Protection Clause of the Fourteenth Amendment whenever it
makes the affluence of the voter or payment of any fee an electoral standard." Harper v. Va. State
Bd. of Elections, 383 U.S. 663, 666 (1966).
182. Initially, Georgia law required that an indigent referred by a nonprofit organization pay
five dollars for photo identification. GA. CODE ANN. § 40-5-103(b) (2004). In 2006, following a
court challenge, Georgia legislators passed a revision of the law that directs the state to distribute the
photo identification for free. GA. CODE ANN. §§ 21-2-417.1, 40-5-103(d) (Supp. 2006); see Common Cause/Ga., 406 F. Supp. 2d at 1369-70 (quoting the affidavit of indigency that Georgians must
sign to obtain a free photo identification); see also IND. CODE ANN. § 3-11.7-5-2.5(c) (West 2006)
(allowing indigent individuals unable to afford a photo identification the ability to cast a ballot).
Michigan Law Review
[Vol. 105:631
pay for a card to vote."' The court concluded that the affidavit likely constituted a "material requirement" imposed solely upon those who do not pay a
fee for a photo-identification card, and thus fell short of compliance with the
Twenty-Fourth Amendment. 1 While the court's conclusion may be correct,
empirical data on the number of voters likely to complete the affidavit of
indigency is more definitive than speculation about the embarrassment and
veracity of voters.
A state might also distribute free photo identification to anyone who
asks without requiring that individuals declare indigency.' 5 As mentioned
above, however, after 2007 the Real ID Act will prohibit states from issuing
photo identification without documentary proof of an applicant's full legal
name, date of birth, Social Security number, and citizenship. Depending on
the state, a birth certificate costs between $10 and $45. A passport costs $97
and certified naturalization papers cost $19.95.1 6 Empirical data would reveal the percentage of the population that lacks ready access to these forms
of documentation and would have to purchase them to obtain a state-issued
photo-identification card to vote.
C. Abridging Voting Rights along Racial Lines
Data is essential in determining whether photo-identification requirements disproportionately dilute the voting rights of people of color.
Congress designed Section 2 of the Voting Rights Act to enforce the Fifteenth Amendment's prohibition on racial discrimination in voting. The
section provides that no voting procedure shall be imposed that "results in a
denial or abridgement of the right of any citizen of the United States to vote
on account of race or color." 8 1 While the vast majority of Section 2 cases
have featured vote dilution challenges to election district boundaries, Sec-
tion 2 also applies to challenges to election practices that disproportionately
deny voting rights to people of color."'
183.
Common Cause/Ga., 406 E Supp. 2d at 1369-70.
184. Id. (citing Harman v. Forssenius, 380 U.S. 528, 542 (1965) (holding that a Virginia requirement that those who do not pay a poll tax must file a certificate of residency constituted a
"material requirement" that abridged the right to vote in violation of the Twenty-Fourth Amendment)).
185.
See GA. CODE ANN. §§ 21-2-417.1, 40-5-103(d) (Supp. 2006); COMM'N ON FED. ELEC-
TION REFORM, supra note 4, at 10.
186. BRENNAN CTR. FOR JUST. & OVERTON, supra note 12, at 4; Application for a U.S. Passport, http://travel.state.gov/pdf/DS-0011 .pdf (last visited Oct. 23, 2006). Bureaucracy may also pose
significant hurdles. See Weinschenk v. Missouri, No. SC88039, 2006 Mo. LEXIS 122, at *15 (Mo.
Nov. 7, 2006) ("In Missouri, the waiting period for a birth certificate alone is six to eight weeks. In
Louisiana, the birthplace of many Katrina refugees who have taken shelter in Missouri, the processing period is eight to ten weeks.").
187.
42 U.S.C. § 1973(a) (West 2000).
188. See Thornburg v. Gingles, 478 U.S. 30, 45 n.10 (1986) ("Section 2 prohibits all forms of
voting discrimination, not just vote dilution." (citing S. REP. No. 97-417, at 30 (1982))); Ellen Katz
with Margaret Aisenbrey, Anna Baldwin, Emma Cheuse & Anna Weisbrodt, Documenting Discrimination in Voting: JudicialFindings Under Section 2 of the Voting Rights Act Since 1982, 39 U.
MICH. J.L. REFORM (forthcoming 2006) (showing that only forty-one of 331 Section 2 lawsuits
Voter Identification
February 20071
A violation of Section 2 is established if:
[Biased on the totality of circumstances, it is shown that the political processes ... are not equally open to participation by ... [voters of color in
that they] have less opportunity than other members of the electorate to
participate in the political process and to elect representatives of their
choice. 9
Plaintiffs need not show that the challenged electoral practice was
adopted with the "intent to discriminate against minority voters," but simply
must show that "as a result of the challenged practice or structure plaintiffs
do not have an equal opportunity to participate in the political processes."' 9
Racial disparities in driver's license and state photo-identification applications are important evidence that a photo-identification requirement to
vote will have a discriminatory impact, but so is data on racial disparities in
how voters establish their identity at the polls. Are voters of color more
likely to use an affidavit, for example, in states that provide that option?
Data on racial disparities in photo-identification possession and use at the
polls from other states is relevant,' 9' but litigants should also commission
detailed studies that analyze racial disparities in the state where the voteridentification law is challenged.
A showing of a disparate racial impact of photo identification alone,
however, is insufficient to establish a Section 2 violation. 92 Courts must also
weigh a nonexclusive list of factors, such as the existence of racially polarized voting, the presence of elected officials who are unresponsive to the
needs of minority voters, whether the policy underlying the contested elec-
tion practice is tenuous, and the effects of past discrimination in areas such
as education, employment, and health.' 93 "The essence of a § 2 claim is that
since 1982 challenged election procedures); see, e.g., Miss. State Chapter, Operation Push, Inc. v.
Mabus, 932 F.2d 400, 401-02 (5th Cir. 1991) (finding that Mississippi's dual registration system
violated Section 2 of the Voting Rights Act); United States v. Berks County, 277 F. Supp. 2d 570,
580-81 (E.D. Pa. 2003) (finding that identification requests from Latino voters, hostility of poll
officials to Latino voters, and other factors led to a Section 2 violation). Daniel Tokaji recognizes
the inapplicability of the leading Section 2 case, Thornburg v. Gingles, to election practices, and
proposes a legal test for election practices. See Daniel P. Tokaji, The New Vote Denial: Where Election Reform Meets the Voting Rights Act, 57 S.C. L. REV. 689, 718 (2006) (recommending a Section
2 test for election procedures in which "a prima facie case [could be made] by showing the challenged practice is a 'but for' cause of racial disparity in voting."). Under Prof. Tokaji's test, the
"state or local entity would still have the opportunity to demonstrate that this practice is necessary to
achieve a compelling government interest." Id.
189.
42 U.S.C. § 1973(b) (West 2000).
190. Thornburg, 478 U.S. at 44 (citing S. REP. No. 97-417, at 2, 15-16, 27, 28 (1982)). Cf
Tokaji, supra note 188, at Part IV.A (asserting that unlike vote dilution cases, vote denial cases implicate the value of participation rather than representation, do not present significant concerns
about proportional representation, and allow for simplicity in measuring disparate impact).
191.
See, e.g.,
192.
See Wesley v. Collins, 791 F.2d 1255, 1260-61 (6th Cir. 1986).
PAWASARAT,
supra note 14.
193. Thornburg, 478 U.S. at 44-45 (citing S. REP. No. 97-417, at 28-29 (1982)). Other Senate factors include, but are not limited to, the following:
[Tihe history of voting-related discrimination in the State or political subdivision ... the extent
to which the State or political subdivision has used voting practices or procedures that tend to
Michigan Law Review
[Vol. 105:631
a certain electoral law, practice, or structure interacts with social and historical conditions to cause an inequality in the opportunities9 enjoyed by black
and white voters to elect their preferred representatives."'
Statistical evidence helps establish whether these other relevant factors
exist in a particular state, including the existence of racially polarized voting, disparities in socioeconomic factors such as education and employment,
and whether the amount of voter fraud is so minimal that the justification for
the photo-identification requirement is tenuous. These various factors will
differ from state to state, and thus the legal status of voter-identification
laws may vary. A federal court might find that a photo-identification requirement to vote in Rhode Island, for example, does not constitute a
Section 2 violation if it finds no history of voting practices or procedures
used to disenfranchise voters of color. Yet an identical photo-identification
requirement in Georgia may violate Section 2 because it interacts with
Georgia's unique social and historical conditions to produce unequal oppor-
tunities for voters of color in that state.
D. "Individual Responsibility" in the Context of Democracy
In determining whether photo-identification requirements comply with
constitutional and statutory provisions, some judges may be tempted to ignore data showing that photo-identification requirements would exclude
legitimate voters and instead focus on the "opportunity" of individuals to
obtain a photo-identification card to vote. Photo-identification requirements
do not constitute a "severe burden" on voting, a poll tax, or a Voting Rights
Act violation, a judge might reason, because most people possess a photo-
identification card and anyone can obtain one.195
This perspective does not ask how many legitimate voters will actually
obtain a fee waiver or return home to retrieve their identification, but instead
enhance the opportunity for discrimination against the minority group, such as unusually large
election districts, majority vote requirements, and prohibitions against bullet voting; the exclusion of members of the minority group from candidate slating processes... the use of overt or
subtle racial appeals in political campaigns; and the extent to which members of the minority
group have been elected to public office in the jurisdiction.
Id.
194.
Id.at 47.
195. The focus on individual responsibility is seen in other election contexts, such as language
assistance at the polls, lifetime bans on felon voting, punch card ballots, and laws that allow challengers at the polls. Individuals, the argument goes, have a responsibility to learn English, stay out
of trouble with the law, punch a ballot correctly, and establish their eligibility to poll challengers.
According to this perspective, the fact that some individuals fail to comply with these norms and
that regulations fall the hardest on particular demographic populations is not a problem that necessitates concern. See, e.g., Wesley, 791 E2d at 1262 (finding that a felon disenfranchisement law does
not violate Section 2 of the Voting Rights Act, reasoning that felons are not "disenfranchised because of an immutable characteristic, such as race, but rather because of their conscious decision to
commit a criminal act for which they assume the risks of detention and punishment"); Stewart v.
Blackwell, 356 F. Supp. 2d 791, 808 (N.D. Ohio 2004) (asserting that there was no "'actual' denial
of the right to vote on account of race" through the use of punch card ballots since "[a]ll voters in a
county, regardless of race, use the same voting system to cast a ballot, and no one is denied the
opportunity to cast a valid vote because of their race").
February 2007]
Voter Identification
whether a "fair" process exists that gives an individual the opportunity to
vote. The vision focuses on the guilt and responsibility of the individual
legitimate voter who lacks photo identification, and does not recognize the
decrease in voter turnout as a harm; if an individual voter fails to comply
with a state mandate, the individual rather than the state is at fault. Individual fraud stigmatizes elections, according to this perspective, but reduced
turnout due to a96 photo-identification requirement does not compromise elec1
toral integrity.
Judges who emphasize individual responsibility avoid issues of vote di-
lution.' 97 As seen in one-person, one-vote cases, "'[t]here is more to the right
to vote than the right to mark a piece of paper and drop it in a box or the
right to pull a lever in a voting booth.""
98
While the simple task of bringing
a photo-identification card to the polls may not appear to be an unreasonable
obstacle for an individual voter, judges should examine whether this requirement reduces voter turnout in the aggregate.
The problem with a focus on individual responsibility is that politics involves not simply individual rights but also associational and structural
concerns.' 99 Through associating with others, individual voters create incentives for politicians to respond to their needs.2 ° Voting is "a vehicle for selfdevelopment and identification, and a means' 20for creating alliances and thus
'
a community among individuals so engaged.
Photo-identification requirements that exclude legitimate voters interfere
with the ability of citizens to identify with one another as a political community, create alliances with others of different backgrounds, and use the
vote instrumentally to enact political change. Despite the emphasis on individual responsibility, photo-identification requirements that exclude
legitimate voters dilute the political choices of not only those who are
196. Cf Kimberl6 Williams Crenshaw, Race, Reform, and Retrenchment: Transformation and
Legitimation in Antidiscrimination Law, 101 HARV. L. REV. 1331, 1341-49 (1988) (describing assumptions of expansive and restrictive visions of antidiscrimination law).
197. Cf Heather K. Gerken, Understanding the Right to an Undiluted Vote, 114 HARV. L. REV.
1663, 1666 (2001) (arguing that vote dilution claims cannot be squeezed into the conventional individual-rights framework); Samuel Issacharoff & Richard H. Pildes, Politics as Markets: Partisan Lockups
of the Democratic Process, 50 STAN. L. REV. 643, 648 (1998) (asserting that the judiciary should "invert the focus of constitutional doctrine from the foreground of rights and equality to the background
rules that structure partisan political competition"); Daniel R. Ortiz, From Rights to Arrangements, 32
Loy. L.A. L. REV. 1217, 1218 (1999) (observing that election law's evolution "has led us away from a
largely rights-based, individual-centered view of politics, to a more pragmatic and structural view of
politics as a matter of institutional arrangements").
198. Reynolds v. Sims, 377 U.S. 533, 555 n.29 (quoting South v. Peters, 339 U.S. 276, 279
(1950) (Douglas, J., dissenting)).
199. Cf Samuel Issacharoff & Pamela S. Karlan, Standing and Misunderstanding in Voting
Rights Law, Ill HARv. L. REV. 2276, 2282 n.30 (1998) (asserting that one-person, one-vote cases
like Reynolds "should be viewed as cases about group political power ... rather than purely about
individual rights").
200.
Gerken, supra note 197, at 1678.
201. Ellen D. Katz, Race and the Right to Vote After Rice v. Cayetano, 99 MICH. L. REV. 491,
513 (2000) (citation omitted).
Michigan Law Review
[Vol. 105:631
unable to produce photo identification but also their allies who do produce a
202
photo-identification card.
Voting is also structural to the extent that one believes that ascertaining
the will of the citizenry as a whole is a central purpose of self-government
in a democracy. Individual votes are counted and weighed relative to one
another, and thus a rule that has a disproportionate impact on a particular
demographic group can "fix" an outcome. Granted, many voting requirements may inconvenience voters and lower turnout, but judges should pay
special attention to those that fail to produce empirically quantifiable benefits that outweigh or are at least equivalent to the costs of the regulation.
Photo-identification advocates recognize the structural elements inherent in
the statement that "voters are disfranchised by the counting of improperly
cast ballots or outright fraud" or that a close election could be determined
by fraudulently cast votes. Judges should not ignore questions of democratic structure and skewed results by substituting the "opportunity" of all to
obtain an identification card for a real analysis of the extent to which photoidentification requirements actually diminish turnout.
V. PHOTO-IDENTIFICATION SUPPLEMENTS AND ALTERNATIVES
In order to assess photo-identification requirements, policy-makers and
judges also need data that measures the comparative effectiveness of other
methods of identifying voters in deterring most fraudulent votes but very
few legitimate ones. This Part reviews two groups of alternatives. The first
group maintains photo identification as an absolute requirement to vote, but
attempts to increase access through measures such as free photoidentification cards, mobile photo-identification card distribution programs,
and election day registration. The second set of alternatives provides measures permitting exercise of the franchise by individuals who arrive at the
polls without photo identification. These measures include, for example, the
option of affirming one's identity by signing an affidavit.
A. Supplements That May Enhance Voter Access
Photo-identification advocates have proposed several supplements that
attempt to mitigate or offset access problems while still requiring a photoidentification card as an absolute condition to vote. These advocates simply
assume that the proposals will address all access issues, and thus ignore the
need for an empirical analysis of proposals designed to enhance access. For
202. Gerken, supra note 197, at 1669-70 (distinguishing vote dilution claims from claims
based on conventional individual fights by observing that with regard to voting, "fairness is measured in group terms; an individual's fight rises and falls with the treatment of the group; and the
fight is unindividuated among members of the group"); James Thomas Tucker, Affirmative Action
and [Mis]representation:Part I-Deconstructing the Obstructionist Vision of the Right to Vote, 43
How. L.J. 405, 414 (2000) ("When an electoral scheme systematically prevents the collective exercise of voting fights for particular groups, the individual fight to vote is diminished accordingly.").
203.
FUND, supra note 22, at 8.
February 20071
Voter Identification
example, a recent Developments in the Law in the HarvardLaw Review that
briefly reviewed the federal district court's decision to block Georgia's
photo-identification law stated as follows:
The hurdles that photographic identification proposals face today could
diminish in as few as two election cycles if the states take on more of the
responsibility of educating voters, ensuring greater access to voter identification facilities, and adhering to HAVA requirements such as cleansing of
the voter rolls. These efforts would minimize at once both the severity of
the proposal's disenfranchising effects and any potential for voter fraud."M
This broad statement makes assumptions without providing empirical
data, and thus policy-makers are unable to assess the proposal's plausibility.
How do we know, for example, that hurdles would diminish "in as few as
two election cycles"? What do studies indicate about government's effectiveness in quickly reducing racial disparities in other contexts? What
specific steps must the state take to educate voters and provide access to
voter-identification facilities, and how does one guarantee that these state
efforts will continue into the future? What happens after 2007 when the
REAL ID Act's enhanced requirements of documentary evidence of citizenship, place of birth, and Social Security number make obtaining photo
identification more difficult?
When policy-makers explore supplements to photo-identification requirements designed to increase access, they should demand specific data
about the effectiveness of such supplements in increasing accessespecially if data shows that fraud is minimal relative to the number of legitimate votes that would be expected to be excluded if the requirements
were to be adopted.
1. Free Photo Identification
In 2006, Georgia allowed for individuals who completed a form declaring indigency to obtain a free photo-identification card,2 0 ' and the CarterBaker Commission proposal would give free photo identification to all nondrivers.2 6 As mentioned above, policy-makers should look to data rather
that free photo-identification programs will resolve all
than simply assuming
..
207
access problems. Some individuals will not take advantage of the programs because they do not know of them, do not have the time to apply, are
204.
Developments in the Law, supra note 11, at 1154.
205. GA. CODE ANN. §§ 21-2-417.1, 40-5-103(d) (Supp. 2006) (revising GA. CODE ANN.
§ 40-5-103 (2004), following a court challenge of the former law, which required indigents to pay
five dollars for a photo identification, so that the state is now required to distribute photo identification for free).
206.
COMM'N ON FED. ELECTION REFORM, supra note 4, at 10.
207. See Publius, supra note 11, at 300 ("Although, as discussed, the claim that minority
voters cannot meet such requirements is unsubstantiated, that problem can be easily resolved. For
any individual who does not have a driver's license or other photo identification and who needs to
obtain one to meet this requirement, states should waive the fee their motor vehicle departments
charge for the nondriver's license identification cards they issue.").
Michigan Law Review
[Vol. 105:631
ashamed to admit indigency, or do not have the resources to obtain the supporting documentation necessary to obtain a state-issued photoidentification card under the REAL ID Act.2°8 Others may secure a free
photo-identification card and lose it, have it stolen, or simply forget to bring
it to the polls.
2. Expanded Photo-IdentificationDistributionthrough
Mobile Buses and More Photo-IdentificationOffices
The Carter-Baker Commission proposed that states take "an affirmative
role in reaching out to non-drivers by providing more offices, including mobile ones," to provide photo-identification cards to voters. 2 9 In Georgia, the
state has commissioned a bus to travel through the state and provide photoidentification cards. Data is needed, however, because the effectiveness of
mobile buses and other outreach efforts rests upon the details of implementation, which may vary based on written policies, budget priorities, and the
dedication and competence of politicians and civil servants.
For210 example, an estimated 300,000 adults in Georgia lack a driver's license. In 2005, Georgia had a mobile photo-identification program that
consisted of one bus that traveled to a location for a day or two, and was
available during the middle of the day from 9 a.m. until 3 p.m. 2 1 Acknowledging the shortcomings of using a hand-me-down bus from another agency,
a spokesperson for Georgia Governor Sonny Perdue said, "'We've got to
have.'
start with the resources we've got and can't spend money we don't
,,212
While the bus had the capacity to issue two hundred photo identifications a
day, it issued fewer than five hundred licenses during the last three months
of 2005 .213
208. See Common Cause/Ga. v. Billups, 406 E Supp. 2d 1326, 1363-64 (N.D. Ga. 2005)
(explaining that the Georgia indigency affidavit was insufficient, and describing various classes of
citizens who would remain without photo identification).
209.
COMM'N ON FED. ELECTION REFORM,
supra note 4, at iv.
210. Carlos Campos, Photo ID Bus Gets Little Use, ATLANTA J.-CONST., Dec. 19, 2005, at LB
("The idea [of the mobile photo-identification program] was to bring photo IDs to the estimated
300,000 voting age people who don't have driver's licenses."); see also Matthew S.L. Cate, Photo
ID Bus Rolls into Northwest Georgia, CHATTANOOGA TIMES FREE PRESS, Dec. 21, 2005, at NG4.
211.
Georgia.gov Home Page, DDS Begins Mobile Licensing Tours & Center Reservations
for Photo IDs, http://www.georgia.gov/00/article/0,2086,4802_496141800330,00.html (last visited
Sept. 10, 2006); see also Georgia Department of Driver Services, GLOW Bus Schedule,
http://www.dds.ga.gov/drivers/glowbus.aspx (last visited Sept. 10, 2006) (showing that currently
there are no scheduled GLOW Bus stops).
212. Nancy Badertscher, State Bus Will Roll for Voter IDs, ATLANTA J.-CONsT., Aug. 9, 2005,
at IB (quoting Heather Hedrick, spokeswoman for Georgia Governor Sonny Perdue).
213. Campos, supra note 210; see also Cate, supra note 210; cf.Common Cause/Ga., 406 F.
Supp. 2d at 1363 (asserting that Georgia's mobile bus program was insufficient because it utilized
only one bus for 159 counties, voters lacked notice of when the bus would be in their area, and the
bus was not wheelchair-accessible). The Georgia governor's spokesperson asserted that this relatively low number proved that "the vast, overwhelming majority of people who want to vote in
Georgia already have valid IDs." Campos, supra note 210. Michigan has a relatively robust mobile
identification program, but ten percent of voting-age citizens in Michigan remain with neither
February 20071
Voter Identification
3. ProvisionalBallots Counted When Photo IdentificationPresented
The Georgia and Carter-Baker Commission provisions also allow voters
who do not bring their photo identification to the polls to cast a provisional
ballot, which officials will count if voters present a photo-identification card
to an elections office within two days of the election. In Georgia, officials
presented evidence that in one county, thirteen people without photo identification had voted provisionally and two of them had returned within the
forty-eight-hour period following the election with a photo-identification
card.21 4 More comprehensive evidence is needed, however, to determine how
many legitimate votes will continue to go uncast or uncounted because (1)
voters do not possess photo-identification cards, or (2) voters do not make or
have the time to return to an elections office.
4. Election Day Registration
States that enact a photo-identification requirement could also adopt
election day registration, which allows unregistered, eligible citizens to
show up at the polls on election day, register, and immediately cast a ballot.
While most states require that voters register ten to thirty days before an
election, six states have election day registration and have enjoyed a voter
turnout increase of nine to fourteen percentage points. 5 Some have asserted
that election day registration invites fraud, but these claims might dissipate
if state-issued photo identification were required to vote.216
Election day registration may increase turnout by removing registrationdeadline barriers for all citizens. Unlike free photo identification and similar
programs, however, election day registration is not targeted at easing the
burden on the specific group of voters who lack photo identification.
driver's licenses nor nondriver's photo-identification cards. BRENNAN CTR. FOR JUST. & OVERTON,
supra note 12, at 7 (citing Telephone Conference with Christopher Thomas, Michigan Dir. of Elections (Sept. 21, 2004) (estimating that ninety percent of eligible voters in Michigan possess driver's
licenses or state-issued identification)). Data from 2003 indicates that 90.2% of driving-age persons
in the state of Michigan possess a driver's license. See FED. HI-GWAY ADMIN., supra note 132.
214.
Common Cause/Ga., 406 F. Supp. 2d at 1365.
215. The election day registration states are Idaho, Maine, Minnesota, Montana, New Hampshire, Wisconsin, and Wyoming.
216. Instead of election day registration, a state could adopt universal registration, in which it
affirmatively registers all voters (not unlike federal officials affirmatively attempt to count all citizens during the U.S. Census). In many other nations around the world, registration is the
responsibility of the state rather than individuals or interest groups. The Carter-Baker Commission
Report did not call for universal registration, but it did state that states should "play an active role in
registering as many qualified citizens as possible." COMM'N ON FED. ELECTION REFORM, supra note
4, at 9. Election day registration may be less expensive and more feasible than universal registration,
however, because the U.S. government is not charged with affirmatively registering all voting-age
citizens.
Michigan Law Review
[Vol. 105:631
B. Alternatives ThatAllow Voters Who Lack
Photo Identificationto Cast Ballots
Several methods exist for confirming the identity of voters who arrive at
the polls without photo identification, all of which evoke questions of the
effectiveness of such methods to prevent fraudulent votes but not legitimate
ones. This section walks through the general contours of various alternatives, and calls for data on each so that policy-makers and judges can make
an informed
comparison between them and photo-identification require217
ments.
1. Nonphoto Identification
Rather than making a photo-identification card an absolute requirement
for voting, a state could expand acceptable documentation to include nonphoto identification, such as a utility bill or bank statement. As discussed in
Part I, this is currently the law for all who vote at the polls in ten states, and
for first-time voters who registered by mail in all states.
Many people without photo identification would likely have nonphoto
documentation, but some would not or would forget to bring it to the polls.
The exclusionary impact of this option might be assessed through analyzing
affidavit data in states such as Connecticut, Delaware, Kentucky, and Tennessee, which allow either photo or nonphoto identification to vote, but also
accommodate voters without such documentation by providing an affidavit
exception.
Photo-identification advocates would likely argue that nonphoto documentation allows for more fraud than photo documentation. 211 Statistical
study is needed, however, to establish the extent to which improper impersonation using nonphoto documentation occurs.
2. Requiring Photo Identification at Registration
Rather than at the Polls
Another alternative would require photo identification at registration
rather than at the polls. Photo identification at registration would primarily
enhance access for people who have obtained photo identification but later
fail to bring it to the polls. The restriction might reduce access because it
would prevent those who lack a photo-identification card from registering.
217. One draft study found that signature requirements, nonphoto identification requirements,
and photo-identification requirements with an affidavit safety net have negative turnout effects of 34% compared to the requirement that voters simply state their names, with larger turnout differences
for subgroups like Latinos and African-Americans. See TIMOTHY VERCELLOTTI & DAVID ANDERSON, PROTECTING THE FRANCHISE, OR RESTRICTING rr? THE EFFECTS OF VOTER IDENTIFICATION
REQUIREMENTS ON TURNOUT 13 (unpublished paper prepared for annual meeting of the 2006
American Political Science Association), http://moritzlaw.osu.edu/blogs/tokaji/voter%2Oid%20and
%20tumout%20study.pdf.
218. See Publius, supra note 11, at 288-89 ("[Ilt is obvious that allowing documents without
photographs is not an acceptable security measure for our voter registration and voting process.").
February 20071
Voter Identification
3. Signature Comparison
Most states without documentation requirements currently require that
each voter establish his identity by signing a pollbook. In many states, the
signature at the polls is compared with a photocopy of the signature the
voter provided when he registered. Any assessment of the costs and benefits
of this procedure should consider the extent to which poll workers detect
fraudulent signatures and prevent fraud, and the extent to which poll workers erroneously allege fraud and block access.
4. Affidavits
In affidavit states, a voter who does not provide photo identification may
sign an affidavit attesting to his identity under penalty of perjury. An alternative option would require a voter signing an affidavit to cast a provisional
ballot that election officials would count only after they electronically
matched the affidavit signature against the signature the voter provided during registration. Studies should investigate the extent to which affidavits
mitigate access concerns (bureaucratic mismanagement might hinderS219access
by some voters) and the extent to which affidavits reduce voter fraud.
5. Indelible Ink
In Iraq, voters dipped their thumbs in indelible ink when they cast a ballot. Indelible ink would not prevent voting by persons ineligible to vote who
impersonate a registered voter, but it would prevent multiple voting by these
individuals.
6. GovernmentMaintains DigitalPicture/Biometric/Thumbprint
Government, rather than voters, could bear the burden of identification
by obtaining a photograph, biometric information, or a thumbprint from
citizens when they register to vote. Officials would make this information
available at polls so that poll workers could confirm the identity of those
who lack photo identification by looking at the voter photograph on file (either printed on the voter registration rolls or accessible via laptop computer)
or by verifying the voter's identity through a biometric or thumbprint device.'2 Empirical studies should examine the extent to which these solutions
219. See Adam Cohen, Indians Face Obstacles Between the Reservation and the Ballot Box,
N.Y. TIMES, June 21, 2004, at A18 (observing that in a South Dakota election, some officials failed
to offer affidavits to American Indians without photo-identification cards).
220.
See, e.g., LARRY J. SABATO & GLENN R. SIMPSON, DIRTY LITTLE SECRETS: THE PERSISTENCE OF CORRUPTION IN AMERICAN POLITICS 322 (1996) (proposing that officials obtain a
thumbprint of voters at registration); Hasen, supra note 11, at 969-70 (proposing that officials obtain biometric data at time of registration); Edward B. Foley, Is There a Middle Ground in the Voter
ID Debate?, ELECTION LAW @ MORITZ, Sept. 6, '2005, http://moritzlaw.osu.edu/electionlaw/
comments/2005/050906.php (last visited Sept. 10, 2006) (proposing that officials obtain a picture of
voters at registration).
[Vol. 105:631
Michigan Law Review
would hamper voter registration, and further normative discussion is necessary regarding privacy issues implicated by the proposals.
7. Better Election AdministrationPractices
State election officials could deter fraud by creating a statewide voter
registration database that is regularly updated and compiling statistics on
voter fraud to observe trends and enforcement efforts. Photo-identification
advocates often argue that voting rolls are filled with dead people and voters
who have moved away, and that these inactive voting files facilitate voter
fraud.2' The Help America Vote Act requires that each state develop a single, comprehensive, computerized, statewide voting list that any election
official in the state be able to access at any time. 2 To keep their lists current, states are required to coordinate with state agencies to ensure that
voters who die or lose their right to vote through felony conviction are removed from the list.2 3 Moreover, the states are directed to cull their lists
actively by removing any voter who does not vote in two consecutive general elections for federal office and who fails to respond to a notice of
removal (although "no registrant may be removed solely by reason of a failure to vote" 224). We would need data on how much list cleansing would
purge could erroneously rediminish access, however, as an overinclusive
25
move legitimate voters from voting lists.
State officials should also compile and maintain statistics on charges and
convictions of voter fraud. Such information could identify which tools are
best tailored to prevent voter fraud.
Finally, rather than simply focusing on voters, antifraud measures
should scrutinize government officials and others who manage elections.
Election officials have a far greater opportunity than individual voters to
determine the outcome of an election through fraud, and partisan election
officials often have greater incentives to commit fraud. A program of regular
and unannounced independent audits of polling places, county election
boards, Secretary of State offices, and private vendors should examine voter
registration and polling place procedures, voting machines, vote-tabulation
systems, software, purge processes, and other procedures. Such antifraud
measures pose little risk of discouraging legitimate voter participation and
are less likely than photo-identification requirements to improperly skew
election outcomes.
221.
See Section II.A for a more detailed review of this argument.
222.
42 U.S.C. § 15483(a) (2000 & Supp. 2003).
223.
Id. § 15483(a)(2)(A)(ii).
224.
Id. § 15483(a)(4)(A); see BRENNAN
CTR. FOR JUST. & OVERTON,
supra note 12, at 11.
225. Cf SPENCER OVERTON, STEALING DEMOCRACY: THE NEW POLITICS OF VOTER SUPPRESSION (2006) (noting that in Florida in 2004, a Republican Secretary of State erroneously purged
about 22,000 African-American voters and 2100 former prisoners who had successfully applied for
restoration of their voting rights).
February 2007]
Voter Identification
681
CONCLUSION
Rather than continuing to rely on unsubstantiated factual assumptions,
election law scholars and policy-makers should look to empirical data to
weigh the costs and benefits of various types of election regulations. Existing data suggests that a photo-identification requirement would
disenfranchise twenty million Americans while deterring minimal voter
fraud. Policy-makers should place a moratorium on photo-identification
proposals until they obtain a better empirical understanding of the extent
and nature of voter fraud and the effect of the proposals on access by legitimate voters.
682
Michigan Law Review
[Vol. 105:631